Skip to main content

Full text of "Forest Tree Improvement Forestry Paper 20"

See other formats




Report on the 

FAO/DANIDA Training Course 
on Forest Tree Improvement 
Merida, Venezuela 
January - February 1980 

The designations employed and the presentation 
of material in this publication do not imply the 
expression of any opinion whatsoever on the 
part of the Food and Agriculture Organization 
of the United Nations concerning the legal 
status of any country, territory, city or area or 
of its authorities, or concerning the delimitation 
of its frontiers or boundaries. 

ISBN 92-5-100943-0 

All rights reserved. No part of this publication may be reproduced, 
stored in a retrieval system, or transmitted in any form or by any means, 
electronic, mechanical, photocopying or otherwise, without the prior 
permission of the copyright owner. Applications for such permission, 
with a statement of the purpose and extent of the reproduction, should 
be addressed to the Director, Publications Division, Food and Agriculture 
Organization of the United Nations, Via delle Terme di Caracal I a, 00100 
Rome, Italy. 

> FAO 1M6 

- iii - 


The FAO/DANIDA Training Course on Forest Tree Improvement was held in Venezuela 
from 14 January to 2 February 1980. The Course was organized by the FAO Department of 
Forestry in collaboration with the Government of Venezuela, the University of the Andes, 
and the Institute Latinoamericano de Investigacion y Capacitacion. The Course was financed 
with funds made available by the Danish International Development Agency (DANIDA). It 
was attended by 19 experts from 17 countries of Latin America: Argentina (1), Bolivia (1), 
Brazil (1), Chile (1), Colombia (1), Costa Rica (1), Cuba (1), Dominican Republic (1), 
Ecuador (1), Guatemala (1), Honduras (1), Nicaragua (1), Panama (1), Paraguay (1), 
Peru (1), Uruguay (1) and Venezuela (3). 

The Course consisted of two weeks of lectures and practical demonstrations in 
Merida and one week study trip to the states of Barinas and Monagas in western and 
eastern Venezuela respectively. 

Lectures included the following topics: tree improvement in relation to national 
forest policy; elements and principles of genetics; conservation and rational use of 
forest genetic resources; collection and handling of forest seeds; storage, testing and 
certification of forest seeds; experimental designs; statistical interpretation of test 
results; species and provenance trials; selection and management of seed stands; 
selection of forest trees; vegetative propagation methods; controlled crossing systems 
and designs; establishment and management of seed orchards; progeny trials; genotype/ 
environment interaction; breeding for disease resistance; strategies for tree development 
programmes; economic considerations of forest tree breeding programmes. 

- V - 



1 . Introduction 1 

2. Organization and Conduct of the Course 1 

3. Conclusions 1 

4. Acknowledgements 2 


1. List of lecturers, support personnel and participants 4 

II . Course programme : theory 8 

III. Course programme: study trip 9 

IV. Lectures : 10 

Tree improvement in Relation to National Forest Policy 

and Forest Management (R.L. Will an) 11 

Elements and Principles of Genetics (W.H.G. Barrett) 15 

Principles and Strategies for the Improved Use of 

Forest Genetic Resources (C. Palmberg) 23 

Sampling in Seed Collection (C. Palmberg) 41 

Collection and Handling of Forest Seeds (C. Palmberg 

and G.H. Melchior) 46 

Storage, Testing and Certification of Forest Seeds 

(B. Ditlevsen) 60 

Experimental Designs (B. Ditlevsen) 75 

Statistical Interpretation of Test Results (B. Ditlevsen) 86 

Species and Provenance Trials (R.L. Willan) 103 

Seed Stands (M. Quijada) 112 

Selection and Management of Seed Stands with Special 

Reference to Conifers ( W.H.G. Barrett) 116 

Selection and Management of Seed Stands: Hardwoods 

(C. Palmberg) 122 

Selection of Forest Trees (M. Quijada R. ) 124 

Quantitative Genetics: General Principles and Practical 

Applications in Forest Tree Improvement (B. Ditlevsen) 131 

Vegetative Propagation Methods (M. Quijada R. ) 140 

- vi - 


Controlled Crossing Systems and Designs (B. Ditlevsen) 148 

Seed Orchards (W.H.G. Barrett) 160 

Progeny Trials (M. Quijada) 168 

Genotype Environment Interaction (M. Quijada) 174 

Breeding for Disease Resistance (C . Palmberg) 178 

Economic Considerations in Forest Tree Breeding Programmes 

(B. Ditlevsen) . . . ."7 187 

Planning and Strategies of a Tree Development Programme 

(C. Palmberg, D.K. Paul & R.L. Willan) L 197 

Some Aspects of the Problem of Genetic Improvement of 

Hardwood Species Native to Venezuela (M. Quijada) 213 

V. Study trip notes 

1. Flowering, Seed Production and Artificial Pollination 

in Bombacopsis quinata in Venezuela 216 

2. Plantation Programme in Llanos Orientales, Venezuela 219 

3. Activities and Human Resources needed for the CVG 

Plantation Programme, Uverito 220 

4. Calendar of CVG Activities , Uverito 222 

VI. Country Reports 223 

VII. Bibliography 266 

1. Publications distributed to participants 266 

2 . Other publications of interest 267 

VIII. Certificate of participation for participants 271 

Opening of the Course. C. Palmberg (FAO) ; Dr. P. Rinc6n Gutierrez (Rector, University of 
the Andes); Dr. R. Chalbaud Zerpa (Governor of the State of MSrida) ; and J.R, Corredor 
Trejo (Dean, Dept, of Forestry Science, ULA) . 

Opening of the course. Dr. R. Chalbaud Zerpa (Governor of the State of MSrida) ; 

C. Palmberg (FAO Co-Director of the course); B. Ditlevsen (DANIDA Co-Director of the 

course); A. Luna Lugo (ULA); and W.H.G. Barrett (International Director of the course). 


The FAO/DANIDA Course on Forest Tree Improvement was held in Venezuela from 
14 January to 2 February 1980, at the kind invitation of the Government of Venezuela. 
The Course was organized by the Department of Forestry of FAO with funds made available 
by the Danish International Development Agency (DANIDA). With the technical collabora- 
tion of the University of the Andes in MSrida and the support of the Institute Latino- 
american de Investigacion y Capacitacion, IFLAIC. 

The training course was one in a series as part of the FAO programme for forest 
tree improvement. Earlier courses in the series were held in Denmark in 1966, in Kenya 
in 1973 and in Thailand in 1976. 

The objective of the course held in the Latin American region, and which was 
intended for professional foresters, was to provide participants with a theoretical and 
practical background in present-day forest tree improvement, and to promote contact 
between foresters and forestry institutions in the region. 

Nineteen participants from 17 Latin American countries participated in the training 
course (see Appendix I). 

Dr. W.H.G. Barrett of Argentina was the International Director of the course. 
Dr. M. Quijada R. of the Institute of Silviculture, University of the Andes, acted as 
national Co-Director. Dr. B. Ditlevsen of Denmark acted as the DANIDA Co-Director and 
C. Falmberg was the FAO Co-Director. 


The first two weeks of the course covered both theoretical and practical aspects 
of forest tree improvement. The third week was devoted to a study trip to the states of 
Barinas and Monagas in western and eastern Venezuela respectively. 

Appendixes II and III list the detailed programme of the course. 

This training course benefitted from the experience of earlier courses. 

With respect to earlier courses, this training course placed heavier emphasis on 
experimental designs and statistical evaluation of trials. Tropical hardwood tree improve- 
ment techniques were also included where possible. 

The close relationship between national forestry policy and tree improvement ^ 
grammes was stressed throughout the training course. An attempt was made to give partici 
pants a broad perspective of the tree improvement issue. The various prerequisites of 
justifying a new tree improvement programme such as available land and a solid reforesta- 
tion progranme were also stressed. 

Before the course the instructors prepared written lecture notes which were dis- 
tributed to participants on their arrival in Mfirida. Participants also received certain 
publications and papers of fundamental importance of general interest. The lectures are 
found in Appendix IV. A bibliography, including the publications distributed to partici^ 
pants, is found in Appendix VII. 

Bef9re their arrival in MSrida, participants filled out a questionnaire on the 
current plantation and forest tree improvement situation in their respective countries. 
This information formed the basis of the brief talk by each participant based on the 
"Country Reports" in Appendix VI. 


The course offered expertise which will be applied and extended by participants 
upon their return to their respective countries. This aspect of the course should cer- 
tainly not be under-rated, but the opportunity for 19 participants representing 17 

- 2 - 

countries to meet and exchange information and experiences and discuss common problems 
with experts from other countries was also, and unquestionably, a unique and valuable 

This was a particularly opportune moment for a course of this type because most 
of the Latin American countries are intensifying plantation activities with an eye to 
meeting their wood and fuelwood requirements, and are including the introduction of 
exotic species and selection and improvement programmes in their activities. 

The study trip at the end of the course was highly important and very valuable. 
Participants were given the opportunity to see problems which can occur in real life 
situations and to discuss possible compromises and ways of solving these problems. 

It should be stressed that the success of the course was largely due to the 
interest shown by participants. This made it possible to overcome certain difficulties 
which arose at the outset, such as the range of technical expertise met at the meeting. 
The camaraderie and good humour which were a permanent feature of the workshop helped 
create a propitious, inspiring atmosphere for the activities of the course. 


FAO wishes to express its appreciation to the Government of Denmark, who sponsored 
the course through the Danish International Development Agency, and to the Government of 
Venezuela who offered to host the course. In addition, thanks are due to the Department 
of Forestry Sciences of the University of the Andes, Merida, for its very valuable admini- 
strative and professional support and to the Consejo de Desarrollo Cientlfico Humanistic 
(CDCH) for its financial support; the Institute Latinoamericano de Investigacion y 
Capacitacion, IFLAIC, which made its lecture rooms and libraries available to the course; 
the Corporacion Venezolana de Guayana, CVG, and the Compaftia Nacional de Reforestacion, 
CONAKE, through whose cooperation the study trip was made possible; the Laboratorio 
Nacional de Productos Forestales, LABONAC, and the Training Centre of the University of 
the Andes for their collaboration throughout the course. 

Special thanks are due to the Governor of the State of Mgrida, Dr. Reinaldo 
Chalbaud Zerpa, who contributed substantially to the success of the course through his 
personal interest, ensuring that participants were thoroughly briefed on the traditions 
and customs of Venezuela and of Mirida. 

Finally FAO wishes to thank all lecturing and support staff and expresses its 
appreciation for the interest of participants and also extends thanks to all those who, 
through their interests and efforts, helped to make this a successful training course. 

- 3 - 


1st Row, left to right: M.E. Qulnteros (Paraguay); J.A. Enricci (Argentina); 

A. Gonzalez (Cuba); A. Gomez de Fonseca (Brazil); A. Martinez (Panama); G. Moreno (Chile). 

2nd Row, left to right: J. Morales (Venezuela); A, Zapata (Venezuela); G.H. Raets (IFLAIC) ; 

J. Campos (Venezuela); R.A. Rodriguez (Dominican Republic); M. Quijada (Venezuela); 

R. Miliani (Venezuela); C. Palmberg (FAO); W.H.G. Barrett (Argentina); D.J. Moreno 

(Bolivia); T. Quintini (Venezuela); R. Escudero (Uruguay); B. Ditlevsen (Denmark); 

P.E. Silvade la Maza (Nicaragua); O.V. Anleu (Guatemala); H.E. Carrillo (Peru); 

D. Villalobos (Honduras); R. Valcarcel (Brazil); G.E. Porras (Costa Rica); A. Ramirez 

(Venezuela); C.C. Castillo (Venezuela); 0, Carrero (Venezuela); A. Copete (Colombia); 

C. Linares (Venezuela). 

Institute of Silviculture , University of the Andes, Mgrida. 

- 4 - 

Appendix I 



Dr. Wilfredo E.G. Barrett 
Fiplasto Forestal 
Maipu 942 - Piso 21 
1340 - Buenos Aires 

Dr. Bjerne Ditlevsen 
Strandvejen 863 
DK-2930 Klampenborg 

Dr. Marcelino Quijada R. 
Seccion de GenStica Forestal 
Institute de Silvicultura 
Universidad de Los Andes 

Miss Christel Palmberg, M.F. 

Forest Resources Division, 

Department of Forestry 

FAO, Via delle Terme di Caracal la, 1 

00100 Rome, 


Ing. Herman Finol U. 
Universidad de Los Andes 
Facultad de Ciencias 
Fores tales, Via Chorros de 
Milla, Merida 

Ing. S. Gimenez Fonseca 
Universidad de Los Andes 
Facultad de Ciencias 
Fores tale s, Via Chorros de 
Milla, Merida 

Dr. Oton Holmquist 

Universidad de Los Andes 

Facultad de Ciencias Fores tales, 

Via Chorros de Milla, 



Ing. L. Rodriguez Poveda 

Universidad de Los Andes 

Facultad de Ciencias Fores tales, 

Via Chorros de Milla, 



Dr. G.H. Raets 

Institute Latinoamericano 

de Investigacidn y 

Capacitacion, Apart ado 36 



- 5 - 



Ing Juan Andres Enricci 
Estacion Forestal Trevelin 
Casilla Correo No. 17 
Revel in - Chubut (9203) 


Ing Deimer Jesus Moreno 


Casilla Correo No. 209 

Avaroa No. 637 



Ing Agostinho Gomes da Fonseca 
I.B.D.F./D.Pq. - Ed. 
Palacio do Desenvolvimento 
13 Andar - SBN. 
Brasilia - D.F. 


Ing Alejandro Copete Perdomo 


Institute Nacional de los 

Recursos Naturales Renovables 

Calle 26 No. 13 B-47 



Lie. Anibal Gonzalez Roque 
Seccion de Genetica 
Centro de Investigaciones Forestales 
Calle 174 No. 1723 E/ 17B y 17C 
Siboney - Marianao 


Ing Jaime Narvaez 
Arosemena Tola No. 452 
Urb. Borja Yerovi Sector 32 


Ing Gustavo Moreno Diaz 
Corporacion Nacional Forestal 
Centro de Semi lias 
Casilla 5 


Ing Guillermo Enrique Porras Sandoval 
Direccion General Forestal 
Ministerio de Agricultura y Ganaderia 
San Jose 


Sr. Ramon Agustin Rodriguez R. 
Direccion General Forestal 
Centro de los Heroes 
Apartado Postal 1336 
Santo Domingo 


Sr. Osman Vinicio Anleu 


Institute Nacional Forestal 

7 a Av. 7-00 Zona 13 

Guatemala Ciudad 


Sr. Angel Danilo Villalobos Nuftez 
Seccion Mejoramiento Genetico 
Escuela Nacional de Ciencias 
Forestales, Apdo No. 2 


Lie. Aristides Martinez Montilla 
Seccion de Mejoramento Genfitico 
Direccion Nacional de Recursos 
Naturales Renovables 
Paraiso, Panama S 


Lie. Pedro Eloy Silva de La Maza 
Seccion de Invest igacion Forestal 
Institute Nicaraguense de Recursos 
Naturales y del Ambiente, I.R.E.N.A. 


Ing Martin Eugenio Quinteros Doldan 
Servicio Forestal Nacional 
Centro Forestal Alto Parana 
Cd. Strossner 

- 6 - 


Ing Hugo Edgar Carrillo Vargas 
Direction General Forestal y de Fauna 
Regi6n Agraria IV 
Prolongaci6n Raimondi S/N 


Ing Oswaldo Carrero 

Centre de Invest igaci5n Forestal 

Companla Nacional de Reforestacion 


Apdo. No. 264 

Maturin/Edo. Monagas 

Ing Roberto Miliani 

Estaclon Experimental de Semi-lias 

Fores tales, M.A.R.N.R. 

El Limon 

Maracay/Edo. Aragua 


Ing Rafael Escudero Rodriguez 
Facultad de Agronomla 
Departamento Forestal 
Universidad de la Republica 
Avenida Garz6n 780 

Ing Tomas Quintini 
Sub-Gerencia Forestal 
Corporaci6n Venezolana de 
Guayana, CVG 

Centro Comercial Los Olivos 
Puerto Ordaz/ Edo. Bolfvar 


Lie. Carmen Cecilia Castillo 
Division de Desarrollo Agrfcola, CVG 
Edificio La Estancia, Piso 13 
Chuao, Caracas 

Sr. Jos Campos 
Sub-Gerencia Forestal, CVG 
Centro Comercial Los Olivos 
Puerto Ordaz /Edo. Bolivar 

Ing Jairo Morales 

CONARE, Programa Coloradito 

Apartado 196 

El Tigre/Edo. Anzoategui 

- 7 - 

Bombacopsis quinata, genetic variation in the occurrence of buttresses. 

- 8 - 

- 9 - 

Appendix III 

27 January 

28 January 

29 January 

30 January 

31 January 

1 February 

2 February 

(Study trips) 

Trip by bus to Barrancas. "El Irel 11 experimental station. 
Controlled pollination demonstration on Bombacopsis Quinata. 

Caemital experimental forest: trials at "El Irel station". 

Trip to Puerto Ordaz, "El Merei" forestry site. 

Visit to CVG plantations, nursery and trials in Uverito. 

Visit to CONARE plantations, nursery and trials in Chaguaramas. 

Trip to Puerto Ordaz, visit to "La Llovizna" park and SIDOR plant 
(Siderurgica del Orinoco). 

End of course. 

"Perils" during the study trip 

- 10 - 

Appendix IV 



Improvement of forest trees in relation to national forestry policy and 

management of forest trees (R.L. Willan) 11 

Elements and principles of genetics (W.H.G. Barrett) 15 

Principles and strategies for the improved use of forest genetic 

resources (C. Palmberg) 23 

Sampling in seed collection (C. Palmberg) 41 

Collection and handling of forest seeds (C. Palmberg & C.H. Melchior) 46 

Storage, testing and certification of forest seeds (B. Ditlevsen) 60 

Experimental designs (B. Ditlevsen) 75 

Statistical Interpretation of trial results (B. Ditlevsen) 86 

Species and provenance trials (R.L, Willan) 103 

Seed stands (M. Quijada) 112 

Selection and management of seed stands with special 

reference to conifers (W.H.G. Barrett) 116 

Selection and management of seed stands: hardwoods (C. Palmberg) 122 

Selection of forest trees (M. Quijada) 124 

Quantitative genetics: general principles and practical 

application in forest tree Improvement (B* Ditlevsen) 131 

Methods of vegetative propagation (M. Quijada) 140 

Controlled crossing systems and designs (B. Ditlevsen) 148 

Seed orchards (W.H.G. Barrett) 160 

Progeny trials (M. Quijada) 168 

Genotype /environment Interaction (M. Quijada) 174 

Breeding for disease resistance (C. Palmberg) 178 

Economic considerations on forest tree Improvement programmes (B. Ditlevsen) ... 187 

Planning and strategies of a tree improvement programme 

(C. Palmberg, D.K. Paul & R.L. Willan) 197 

Some aspects of the problem of genetic Improvement of 

hardwood species native to Venezuela (M. Quijada) 213 

- 11 - 

Tree Improvement in Relation to 
National Forest Policy and Forest Management 

R.L. Willan 


Introduction 11 

Prerequisites for a tree improvement programme 12 

Defining objectives 12 

Constraints 13 

Tree Improvement in relation to management 13 

Summary 14 

Bibliography 14 


The justification of any national programme of tree improvement and the amount of 
resources devoted to it should be largely determined by the forest policy of the country 
(Keiding 1974). Forest policy, in turn, must be related to national development plans. 

Basic questions which need to be asked are (1) How significant is forest in the 
nation's economy? and (2) How significant is plantation forestry in the forestry sector 
as a whole? At one extreme is the case of a small, highly populated country in which 
there is no room for extensive forestry, and therefore no place for tree improvement. 
Singapore and Malta are examples. At the other extreme are countries with low population 
density and large areas of natural forest capable of being naturally regenerated and of 
providing the country's needs in the foreseeable future. Gabon and, until recently, 
Kalimantan and Indonesia may be examples. In such countries there is a case for a simple 
silvicultural technique designed to retain a proportion of the best trees as seed-bearers 
and, without doubt, for the establishment and protection of strict natural reserves 
within the forest, but not for a long-term programme of tree breeding. In regions where 
the crucial protective function of the forest in steep topography precludes commercial 
harvesting and hence reforestation, there is likewise no scope for tree breeding. 

Few countries can now afford to rely entirely on natural forests. In most the 
pressure on land is becoming steadily greater and, if the requisite raw materials can be 
produced from a much smaller area or in a shorter period by using intensive high-yielding 
plantation methods, foresters have a responsibility to employ these methods. Concentrated 
plantations lead to indirect economic and social benefits through the development of forest 
industries and opportunities for employment. Accordingly, plantations are being established 
on an increasing scale throughout the world. In practice the roles of natural and man-made 
forests should be considered as complementary, with natural forests fulfilling essential 
protective and cultural functions and providing certain special categories of wood, such 
as high quality material for cabinet making or veneering, while plantations supply an in- 
creasing proportion of requirements for constructional timber, utility sawn wood, wood 
based panels and pulp (Hughes and Willan 1976). 

Generally speaking, the prerequisite for a tree improvement programme is plantation 
forestry. As soon as seed is collected, plants raised and artificially cultivated, there 
is a chance to select and improve. Thus, it will be convenient to start the consideration 
of tree improvement in relation to afforestation and reforestation programmes (Keiding 
1974). This is why country statements which participants have been requested to complete 
set the information on national tree improvement programmes in the context of the national 
environment, the national forest policy and national programmes on afforestation and re- 

- 12 - 


Prima facie Even though there may be a case for plantation forestry and tree breed- 
ing in a country, expenditure of funds and effort demands certain prerequisites :- 

A. The planting programme 

(1) Availability and control of land. The large investment involved in planta- 
tion forestry can be justified only if there is an assurance that forestry will remain the 
object of land management for at least one rotation. Even with "fast-growing species" 
this is likely to extend for one or more decades. And the managerial authority must have 
full control of the land throughout the period. Excellent security of tenure may exist 

if the plantations are on government-owned forest land included in a national land usage 
plan by which continuity of management is promised for some years ahead. Land which is 
fragmented among numerous small private owners is usually unsuitable for plantation fores- 
try. On the other hand, tree breeding may have a part to play in the provision of trees 
for diffuse planting in agrisilviculture, provided that the farmer is convinced of the 
value of the product and of the need to protect and manage the trees. 

(2) Scale of operations. No matter how great the gains to be derived from tree- 
breeding, the basic minimum costs of a small research unit need to be spread over an 
adequate area if they are to pay for themselves. For example a unit costing $100 000 a 
year and producing improved seed capable of yielding an increase in discounted product 
value of $100 per ha per year, would more than pay for itself on a 10 000 ha a year 
programme but could not be justified for a 100 ha a year programme. 

(3) Availability of markets. There needs to be reasonable assurance of markets 
for plantation produce, either within the country or through export. Not only must the 
markets exist, but they must be within economic distance. Plantations, even on* high- 
yielding sites, may be uneconomic if transport costs are crippling. 

B. The tree improvement programme 

(4) There must be reasonable assurance, e.g. in a written policy statement by the 
financing authority, that staff and funds will be provided to the tree improvement pro- 
gramme on a continuing basis* As Zobel (1969) has stated: "Will I have the backing in 
funds, facilities and manpower to do a decent job? If not, then don't start! A half- 
hearted programme, poorly done, will only sour people on forestry and its potentials." 

If a trained tree breeder is not already available from within the country, provision to 
train one must be made from the start. 

(5) Assessment of technical information available from elsewhere. Results of 
research in other countries with similar environments may reduce, if not eliminate, the 
need to start a national tree improvement programme de novo. For small countries with 
modest planting programmes, a regional research unit may provide the same results as 
several national programmes and at a lower cost. Examples of regional tree improvement 
programmes are those which operated in East Africa during the 1960s and ' 70s and CATIE 
now operating for Central America. Even where a country's planting programme is large 
and is carried out on a unique range of sites, International exchange of information and 
of genetic material can do much to avoid duplication and concentrate research on solving 
the most important problems or exploiting the most promising opportunities. 


Having established that the basis for a national tree improvement programme exists, 
the policy maker needs to define its objective as clearly as possible. There are often 
good reasons for specifying more than one objective. In view of the long-term nature of 
forest crops and, in comparison, the speed with which technological preferences and markets 
can change, there is often a strong case for maintaining the maximum flexibility in the 
objectives. Flexibility is especially important with rotations longer than 15 years 
(Hughes and Willan 1976). However, there are limits to the flexibility of programmes 
and species which should be recognized from the start. A high density eucalypt is likely 
to be superior to a low density eucalypt for fuelwood but inferior for short-fibre pulp- 
wood. In the southern USA separate seed orchards are planted for high-density and 

- 13 - 

low-density wood in pines. Thus different purposes may sometimes be served best by grow- 
ing separate crops using separate selection criteria rather than by attempting to grow 
a "multi-purpose 11 crop which serves neither purpose adequately. If more than one objec- 
tive is specified and in case resources prove inadequate to achieve all objectives - which 
is almost inevitable - it is important to assign relative priorities. 

The objectives defined must be based upon immediate, short term and long term 
requirements of the relevant afforestation programme. Constraints on the achievement of 
objectives also must be taken into account at the time of definition, and should indicate 
the limits of the resources (staff, facilities and finance) within which the project will 
be required to operate (Hughes and Willan 1976). 


Tree improvement objectives can only be achieved within existing biological human 
and financial constraints. In some cases these are so limiting that the objectives are 
impossible of attainment and it becomes the duty of the tree breeder to convince the 
policy maker to set more realistic objectives. Examples in which limiting conditions 
could force a change in the objectives might be:- 

(1) Attempts to grow high-yielding plantations in semi-arid conditions. Species 
of provenances may be fast-growing or drought-resistant, but rarely both. Any improve- 
ment from tree breeding is likely to be small and slow. More impressive improvement may 
be obtained by shifting the afforestation scheme to a more humid zone or by introducing 

(2) Attempts to start a tree improvement programme in an introduced species, 
using small plantations of narrow or unknown genetic base. In such cases further intro- 
ductions and provenance trials are more urgent than individual selection in the existing 

(3) Attempts to use an exotic species of which foreign seed supplies are inade- 
quate and which fails to produce a seed crop in the new environment. 

(4) Attempts to improve traits which are known to have very low heritabilities. 

(5) The time-scale imposed by the period taken from germination to seed production 
in forest trees is often a serious constraint. If somewhat improved seed is needed in 
five years and the species takes 10 years to produce seed in a seed orchard, then estab- 
lishment of a seed orchard now cannot achieve the objective. Alternatives, such as heavy 
improvement fellings in commercial stands, or a change of strategy, such as development 

of vegetative propagation methods, must be considered. 


Tree improvement is simply one tool available to forest management and cannot be 
considered in isolation. Research on tree improvement should therefore be closely inte-* 
grated with other research, e.g. on soils and site assessment, on establishment techniques, 
on spacing and thinning and on wood quality. National forest policy must therefore in- 
clude provision for the integration of all sectors of forest research and for the rapid 
multiplication and introduction into forestry practice of improved material. 

Tree improvement results may modify forest management. For example better trees, 
more uniform and more disease resistant, make possible a wider initial spacing which 
reduces costs while still achieving the desired final stocking (Zobel 1969). Breeders 
can produce genotypes which respond well to intensive site preparation and fertilization 
and other genotypes which will tolerate poor soils and no fertilizers. Within a single 
country there are likely to be a number of different site types to be planted and the tree 
breeder therefore needs to develop a range of genotypes adapted to the various sites. 
Technological developments may alter breeding priorities, e.g. the invention of a new 
pruning device leading to a major reduction in pruning costs may reduce the need to breed 
for small branches and early self-pruning, or a revolutionary new pulping process might 
reduce the Importance of wood quality or wood uniformity. 

- 14 - 


(1) Tree improvement programmes must be closely related to the objectives and 
priorities of national afforestation programmes and hence to the national forest and 
development policy of the country. 

(2) A number of prerequisites, e.g. availability of land and a substantial scale 
of afforestation, are prerequisites to justify the initiation of a tree improvement pro- 

(3) If a tree improvement programme is to be started and is to have a good chance 
of success, it must have clearly defined objectives and priorities, and some assurance of 
continuing provision of adequate resources to achieve them. As far as possible objectives 
should be unequivocal and quantitative. 

(4) The tree breeder must examine objectives closely in the light of biological, 
human or financial constraints and, if necessary, he must convince policy makers to change 
the objectives. Only after realistic objectives have been set and accepted can the tree 
breeder start to plan his strategies and programme as described in another lecture. 

(5) Tree improvement interacts with other aspects of research and management, 
e.g. site assessment and preparation, spacing and thinning. So tree improvement research 
must be closely integrated with other sectors of the national forest research programme. 

(6) Tree improvement programmes should be flexible and provision should be made 
for regular revision. National economic priorities change and may call for changes in 
tree improvement objectives. Equally, successes and failures in tree improvement 
research may suggest changes in forest management and afforestation policy (change of 
species, ability to afforest new sites hitherto considered uneconomic to plant). 
Communication between policy and research must be two way. 


Hughes, J.F. & Willan, R.L. 

Keiding, H. 

Zobel, B. 

Polltica, planificaci6n y objetivos. En; Manual sobre 
Investigaciones de Especies y Procedencias con Referencia 
Especial a los Tr6picos. Tropical For. Paper No. 10 & 
lOa. Oxford, UK. 

Tree improvement in relation to national forest policy. 
In report of FAO/DANIDA Training Course on Forest Tree 
Improvement, Kenya. FAO Rome. 

A tree improvement program for a developing country and 
the effects of tree improvement on forest management. 
_In FAO - North Carolina State Forest Tree Improvement 
Training Centre Lecture Notes, Raleigh, USA. 

- 15 - 


W.H. Barrett 
Fiplasto S.A., Buenos Aires, Argentina 



Chromosome structure 

Gene structure 


Non-genie inheritance 17 

Genotype and phenotype 17 

Cell division and new gene combinations 18 

Fertilization and seed development 18 

Endogamy 18 

Hybrid vigour 21 

Bibliography 21 


Genetics is the study of the similarities and differences between individuals, or, 
in other words, of variation and inheritance in living things and consequently their 
evolutionary process. A distinction may be made between conventional or Mendelian (or 
transmission) genetics, and population genetics (the specific effects of genes on living 
things). Genetic improvement is the application of the science of genetics for the good 
of humanity. 

This article describes some basic concepts of genetics, which are applicable both 
to animals and plants. Some examples are given of their application to tree species. 


All living cells are made up of an outer cell wall, cytoplasm and nucleus. Within 
the nucleus are the chromosomes, which are the carriers of genetic information and are 
responsible for transmitting this information to the other cells. They are constant in 
number for each species. Thus, for example, willow and poplar species have n-19 chromo- 
somes, pines n-12, eucalyptus n-11, etc. Cell size varies according to the species, 
usually small with angiospers (which are only a few microns long) and larger with 

I/ This paper is based on: 

- Wright, J.W. (1962). Genetics of Forest Tree Improvement. Forestry and Forest 
Products studies No. 16, FAO, Fome; 

- Wright, J.W. (1976). Introduction to Forest Genetics. Academic Press. 

- 16 - 

The chromosome is a long thread-like structure consisting of deoxyribonucleic acid 
(DNA) and a protein sheath. DNA, the active genetic material, is a long molecule composed 
of two spiral strands. Each strand is composed of four organic bases (cytosine, guanine, 
adenine and thymine) and attached sugar radicals; the two strands are joined together by 
phosphate radicals. This group generates a molecule from a nucleotide. The two strands 
are held together more loosely by hydrogen atoms forming molecules of nucleic acid. 

The ability of DNA to replicate itself makes it possible for the chromosomes to 
transmit genetic information from one generation to the next. This is a consequence of 
the double-strand nature of the molecule and of the properties of the four protein bases. 
Adenine and thymine (purinea with two rings) are joined together by two hydrogen bonds, 
and guanine and cytosine (pyrimidines with one ring) by three hydrogen bonds. For this 
reason, guanine will always be linked with cytosine, and thymine with adenine. At the 
time of chromosome division the two strands unravel , and are replicated, and pair off in 
such a way that each protein base is linked with its complementary base (adenine with 
thymine and guanine with cytosine). To give an idea of dimensions the distance covered 
by ten nucleotides on one strand is 34 8. The bases are arranged linearly along a single 
strand in groups of three, and, as there are four bases, there are 64 combinations. In 
other words, they are like a four-letter alphabet with 64 three-letter words, which are 
organized in special sequences or paragraphs which direct growth processes. The sequences 
responsible for the formation of some amino-acids and for the activity of certain enzymes 
governing protein synthesis are known for some one-celled plants. The molecular weight of 
the DNA molecule is very heavy. It has been estimated that a single pine cell contains 
some 50 000 000 000 nucleotide pairs. If each nucleotide is considered a letter of the 
alphabet, some 2 500 000 pages could be written. 

The control of growth processes of a cell is accomplished by a substance similar to 
DNA called ribonucleic acid (RNA), which differs structurally from DNA inasmuch as it is 
single-stranded. RNA sugars have one more oxygen atom (ribose) and uracil replaces thymine. 
RNA performs the function of carrying messages between the DNA and the other parts of the 
cell, and regulates and participates in the synthesis of amino-acids, and in the synthesis 
of proteins from amino-acids. 


From a structural point of view, a gene is defined as a sequence of triplets along 
a DNA molecule. However, so far it has not been possible to isolate and study the struc- 
ture of a gene. A gene can also be defined as that part of a chromosome responsible for 
the development of a particular trait. Thus we may speak of genes for rapid growth, 
resistance to cold, leaf, length, etc. The gene is considered as the ultimate hereditary 
unit, although we know it is of large molecular size. The number of genes in an indivi- 
dual cell is unknown, although there are estimates for some plants, such as, for example, 
the Pinus banks! ana, in which there are said to be 13 000 000 genes. In some agricultural 
crops it has been found that individuals of the same species can differ by 500 genes. For 
an improvement programme, a few dozen genes could be used. The genes can have large or 
small effects. For example, growth rate is usually governed by many genes with small 
individual effects. On the other hand, the colour of eyes in human beings is governed by 
a single pair of genes with relatively large effects. Some genes may act independently; 
some, however, may act only when others are present. 

Genes are arranged linearly on the chromosomes. For this reason, the genes of one 
chromosome belong to the same linkage group. However, this linkage is not perfect. At 
meiosis, when homologous chromosomes pair, they break and exchange parts. When this 
happens, it is said that there has been a "cross-over 11 , giving rise to new genetic combi- 
nations. Usually there are one or more cross-overs per chromosome. Logically, the 
probability that two linked genes will separate from each other is proportionate to the 
distance between them on the same chromosome. This fact is used to prepare chromosome 
maps, where the distance between genes is defined in terms of frequency of new combina- 
tions; one cross-over unit is equal to 1 percent new combinations. Despite the fact that 
gene linkage has not yet been measured in trees, a knowledge of it is Important for tree 
breeding. For example, in pines there are 12 pairs of chromosomes and therefore 12 
linkage groups. When selecting for any one character, there are apt to be changes in the 
frequency of other, closely linked genes. 

- 17 - 

Genes occupying the same locus in a pair of chromosomes are called "alleles" or 
"allelomorphs" and are denoted by a single letter. Dominant allelic genes cause a 
characteristic to be expressed even when the tree is heterozygous, whereas recessive genes 
only have a visible effect if the tree is homozygous for them. This type of dominance is 
complete. It is partial or incomplete when the characteristic is intermediary in expres- 
sion. This is what happens with red flowers (AA), white flowers and pind flowers (Aa). 
When dominance occurs between non-allelic genes it is called epistatsis. 

Genes are called additive when they enhance each other's effects in a cumulative 
manner. They are genes with small effects, that govern a single trait. 


An error in the replicating process of chromosomes produces changes which are 
called mutations. These usually occur at the level of the genes; they can cause a gross 
change in chromosomes, either in number, structure, inversion, addition or reduction in 
size or replication, etc. This produces large effects, usually abnormalities most of 
which are harmful. 

The frequency of gene mutations is estimated at between 1 in 10 000 and 1 in 
100 000 000. The majority of these mutations are harmful and recessive. Despite that, 
it is a beneficial process, since it is the principal source of variation of living things. 
These changes, because they are recessive, are incapable of expressing themselves. How- 
ever, genes which may be detrimental in one environment may be beneficial in another en- 
vironment. Gene mutation can be increased artificially through the use of X-rays, 
chemical substances, etc. These chemical substances can act directly on the DNA, 
producing chemical changes which, during replication, produce mutant offspring. This 
happens with nitrous acid or ethanosulphonate ethyl. On the other hand, other substances, 
like 5-bromo uracil, only act during the DNA synthesis. By analogy of bases, it resembles 
and replaces thymine in replication. During replication this base can occasionally be 
coupled with guanine instead of adenine, producing a different triplet. 


Cytoplasmic or maternal inheritance has not yet been observed in trees, although 
it has in herbaceous plants. In maize a plant was found which did not produce masculine 
flowers and was therefore sterile, due to a cytoplasmic factor. This trait was considered 
very useful for the production of hybrid maize. Later it was discovered that the same 
clone was very susceptible to disease, and it was therefore rejected. 

There have been cases of paternal inheritance with a Japanese conifer, Cryptomeria 


Genotype is defined as the genetic constitution of an individual or groups of 
individuals with a similar gene make-up as compared with specific genes. 

Phenotype is the external appearance, partially controlled by the genotype. How- 
ever, there are recessive genes that are expressed due to the presence of dominant genes, 
though they do exist in the genotype. There are also genes with small effects, or modi- 
fying genes, that are not manifest in the phenotype. Phenotype is also partly determined 
by the environment. A genotype can contain many genes for rapid growth, without this 
being apparent, because it is on poor soil or in an adverse climate. Or it may have 
genes susceptible to a disease, and appear resistant because the disease is not found 
at that site. However, some characteristics of the genotype may be deduced from a careful 
study of the phenotype. Of course, much more information will be obtained from studying 
a tree's offspring and the relationships between the characteristics of the parents and 
those of the offspring. 

- 18 - 


The cell division that occurs in the cambium, root, tips, leaves and other growth 
mer is terns, is called mitosis* The regular constitution of the cells (2n-diploid) is 
maintained by replicating each chromosome constituting two Identical new chromosomes. 
Here there is no difference between the chromosomes, apart from possible mutations. 

Conversely, in meiosis, there is a reduction In the chromosome number from 2n in 
vegetative cells (diploids) to In (haploid) in the gametes. During meiosis, the two sets 
of homologous chromosomes come together in the centre of the mother cell, and pair there. 
This is where cross-overs occur. Segments of each pair of chromosomes are Interchanged, 
producing new gene combinations. Meiosis is a two-stage process, in which there are two 
successive nuclear divisions that end by forming four gametes. 


In fertilization, the male gamete (n) comes together with the female gamete (n), 
to produce the egg cell (2n) which through successive divisions forms the embryo, the 
seed and later the tree. During this process, chromosomes from the male gamete are in- 
troduced into the nucleus of the egg cell. Meiosis and fertilization are the mechanisms 
whereby genetic variability makes it possible for individuals to segregate and re combine 
and produce offspring genetically different from the parents. 

Anglosperms are characterized by dual fertilization; i.e. one nucleus of the male 
gamete fertilizes two polar nuclei making up the endosperm. As a result, various geneti- 
cally different tissues may be found in the seeds of the angiosperms. 

Tissues Number of chromosomes Origin 

Seed Coat 2n maternal 

Embryo 2n maternal and paternal 

Endosperm 3n 2/3 maternal 1/3 paternal 

In gymnosperms, however, the endosperm is haploid and of maternal origin. 


Endogamy of inbreeding is the crossing of related individuals with loss of hetero- 
zygosity. The extreme case is "self ing", or the crossing of a tree with Itself. There 
are all gradations from this extreme, for example, back-crossing with one of the parents, 
crossing between siblings, cousins, etc., or among members of small Isolated populations. 

Wild populations (forests where there has been no human activity) usually carry a 
large "genetic load" with a considerable number of recessive genes. Since the Individuals 
cross with each other, and not all of them have the same recesslves, these recessive 
traits do not appear in their offspring. However, if the tree is selfed, for each hetero- 
zygous character 25 percent of individuals among the offspring will show the effect of 
this recessive gene. Aa x Aa - 25% AA + 50% Aa + 25% aa. It can happen that the effects 
of these harmful recessive genes are small and not visibly obvious, but if the selfed 
parent contains 100 recessive genes their cumulative effect on the offspring can be 
important. Experiments made with Pinus taeda in the United States showed a loss of 
vigour among offspring of selfed trees as compared with open-pollinated offspring of the 
same trees. With other forest species, selfed seedlings have often grown less than 50 
percent as fast as normal seedlings. 

However, self ing does not always bring with it a loss of vigour. Some autogamous 
plants, such as tomato and wheat, have flower structures that promote self ing. These 
plants may have contained, or acquired through mutations, harmful genes, which have been 
eliminated by thousands of generations of self ing. As far as may be discovered, loss of 
vigour is not due to self ing or to inbreeding per se, but to accumulations of harmful 
recessive genes* A tree free of these recessive genes could be selfed without loss of 

- 19 - 

Allogatnous plants can have mechanisms that inhibit self-pollination, a factor that 
leads to a reduction in the quantity of seed. One of these mechanisms is the existence of 
genes for self-sterility, like the multiple alleles 8^, 82, 83, 84 etc., where neither $1 
nor 82 pollen can germinate a style with the genotype Si 82 etc. There are 40 known 
alleles in red clover. However, it can happen that the pollen fertilizes the egg, but the 
recessive genes act against the embryo, thus producing a sterile seed. 

There is a wealth of literature on the effects of self ing on tree species. 
J.W. Wright quotes examples of loss of vigour and reduction in seed fertility in eucalyp- 
tus, pine, elms, larches, etc., and others with variable results, ranging from vigorous 
to weak offspring and from high to low seed fertility. 

This extreme endogamy selfing produces individuals or lines scarcely or not at all 
adapted to the surrounding environment. Without reaching this extreme, when the crossing 
occurs between a few trees, some random gene fixations take place producing uniform indi- 
viduals little adapted to the environment. 

To sum up, the general effects of endogamy, fixation of characters, decline in 
fertility, size and vigour, are due to an increase in homozygosity. 

Measuring endogamy: In the special case of selfing, heterozygosity is lost at the 
rate of 50 percent in each generation; F is the endogamy coefficient used to measure the 
loss of heterozygosity or increase of homozygosity. The symbol AF is used to express 
the per generation change in amount of inbreeding. 

A? - 1 


N being number of individuals 

In this case N is the total number of individuals where each tree is used both as 'male 
and female parent. When they are different, the formula to be used is: 

+ N (g) 


To calculate inbreeding over a number of generations, it is better to work with the 
heterozygosity coefficient, H 1 - F. 

self fertilization 

double first cousins 

octuple third cousins 


Figure 1. Percentage of homozygosity in successive generations under inbreeding 

various systems (from 8. Wright, 1921. Systems of Mating, Genetics 6: 167), 

To calculate Hn, n being the number of generations, Hn - (H) n ; and H the 
heterozygosity per generation. As numerical example, assume that N 5 andA F 1 


H (per generation) - 9 ; H 2 - (9)2 - 0.810; H 3 - 

10 (10) (10) 


n - 10 

n 100 













- 20 - 

After three generations, 72.9 percent of the original heterozygosity remains in the 
population, and the inbreeding rate is: 

F 3 - 1 - H 3 - 1 - 0.729 - 0.271 

Inbreeding coefficients of populations maintained as sites of N - 50 to 250 for 
10 or 100 generations. 

Population Coefficient F 







(extract from J.W. Wright, 1976) 

In tree breeding it is assumed that the loss of vigour through unbreedlng is pro- 
portionate to the loss of heterozygosity as estimated by the coefficient F. In annual 
plants, on the contrary, it is easy to take five individuals and reproduce them over 
successive generations and measure the results. 

There are mechanisms in plants that encourage or prevent autogamy. Annual plants 
usually have perfect hermaphrodite flowers and have developed mechanisms suitable for 
self ing. This is not generally the way with trees, except (according to J.W. Wright, 
1976) for tropical trees constituting mixed forests with hundreds of species per hectare, 
where the possibility of crossing is remote and there is therefore a high probability of 
self Ing or at least a high percentage of endogamy. There must therefore be mechanisms 
to prevent this damaging the species. Unfortunately we have no experimental data on 
this subject. 

In temperate or cold climates, however, where tree species tend to grow in stands 
which are pure, or composed of a few species only, there are mechanisms inhibiting self- 
ing. In these cases a greater variability is necessary since there are great fluctuations 
in rainfall, temperature, exposure to fire, glaclatlon, etc. One of these mechanisms is 
the dioecious character of the plants, e.g. male and female flowers are produced on 
different trees. This happens with poplars, willows, junipers, Araucaria araucana, 
A. angustlfolia, etc. Monoeclsm also performs this function, e.g. male and female 
flowers appear on the same tree, as with oaks, Aesculus , and most conifers. In the 
latter group the masculine flowers are produced on the lower branches, and the female 
flowers on the higher ones. As pollination is wind-borne, this stratification inhibits 
self Ing to a large extent, since the male pollen would have to rise. 

There are mechanisms like dichogamy, whether in hermaphrodite or In diclinous 
flowers, where the male flower Is receptive at a different time than the female; for 
example, the hazel tree with diclinous flowers (protandry) and the tulip tree 
(Liriodendron) with hermaphrodite flowers (protogeny) . 

A simple method has been developed by forest geneticists to determine the per- 
centage of self ing in natural populations of different forest species . It consists in 
counting the number of abnormal seedlings from open pollinated seed, in the nurseries 
and multiplying the figure by 5. In various pine species between 2 and 7 percent of 
self ing has been recorded. 

"Self ing plus later crossing" has been used to improve herbaceous plants, but it 
lacks practical possibilities when applied to trees. 

The random fixation of genes in small populations is called genetic drift. If 
inbreeding is continued over a long geological period, the result Is a uniform population 
which hae developed non-adaptive traits. The isolation of small pine populations In the 
mountains of Mexico has produced a great diversity of forms, which is probably due to 
random gene fixation. Although this is more frequent in mountain areas, it has also been 

- 21 - 

observed In existing islands, or regions that were under water in different geological 
periods. This is the case of Pinus elliottii var. densa which has highly differentiated 
morphological and adaptive variations, irrespective of distance, in the absence of 
clonal gradients. One may observe on one island key, for instance, a population which 
has fascicules with two relatively short leaves and, on a neighbouring key, fascicules 
with three long leaves. 

J.W. Wright (1976) observed similar examples in species of Picea, noting that many 
of these populations which have not adapted to the environment in which they live are on 
the point of extinction. 


Hybrid vigour or heterosis refers to the exceptional yield of a hybrid when com- 
pared with its parents. To explain this phenomenon, four explanations have been suggested, 
which are: dominance, over dominance, additive and hybrid habitat hypotheses. 

According to the dominance hypothesis, hybrid vigour is due to the absence of 
depression caused by endogamy. The homozygosity produced by inbreeding leads to the 
appearance of recessive genes which are harmful to the plant, producing a depression 
which would otherwise be "covered" by dominant genes from heterozygous individuals. 

The overdominance hypothesis maintains that the heterozygous gene combination 
produces effects impossible to achieve with homozygous genes. 

The additive hypothesis holds that the trait is governed by various genes which 
hybrids accumulate, the effect being thereby greater. 

In the "hybrid habitat" hypothesis, vigour derives from a "hybrid" environment, 
where the parents show less capacity for adaptation than the hybrid. 

If the first hypothesis, of dominance, is true, it should be possible to select 
against recessive genes in subsequent generations and thus fix hybrid vigour. If it is 
due to overdominance, vigour will be greater in the first generation and decrease in 
successive ones. If it is due to additive effects, it should be possible to select the 
best genes governing the components of this vigour, and to obtain a greater effect in 
successive generations. Lastly, if the hybrid habitat hypothesis is true, hybrid vigour 
will be equal in the generations following the F]_ generation. 

A case of extensive utilization of hybrid vigour was the crossing of Pinus rigida 
with . taeda in South Korea, where performance was excellent in F^ and even better in F2- 

The breeding method depends on the type of inheritance of the trait to be bred 
for. If greater productivity is due to dominance or additive effects, selection should 
exclude self ing. If, on the other hand, vigour is due to over dominance "self ing and 
later crossing" is the best method. 


Beadle G.W. and M. Beadle, 1971. Introducci6n a la nueva Genetica (translated from 
English). Editorial Universitaria de Buenos Aires, Argentina, 282 p. 

Herkowitz I.H., 1965. Genetics (2nd ed.), Little, Brown and Co., Boston, 554 p. 
(Spanish translation, GenStica D.E.C.S.A. Mexico, 765 p. 1970). 

Lacadena J.R., 1976. Gengtica (2nd ed.), A.G.E.S.A., Madrid, 972 p. 

SRB, A.M., R.D. Owen and R.S. Edgar, 1965. General Genetics, Freeman and Co., San 
Francisco (Spanish translation, Ed. Omega, Barcelona, 632 p. 1974). 

Strickberger M.W., 1976, Genetics (2nd ed.), Macmillan Publ. Co. Inc., N.Y. 914 p. 
(Spanish translation, 2nd ed. Omega, Barcelona 1978). 

- 22 - 

Wright, J.W. 1976. Introduction to forest genetics. Academic Press, 463 p. 

Wright, J.W. 1962. Genetics of Forest Tree Improvement. Forestry and Forest Products 
Studies, No. 16, FAO, Rome, 399 p. 

Participants in class. 

- 23 - 


C. Palmberg 

Forest Resources Division 

Forestry Department 




Introduction 24 

Principles of conservation and utilization of genetic resources of forest trees . 24 

Exploration 25 

Collection for Evaluation 25 

Evaluation 25 

Conservation 26 

In situ 26 

Ex situ 27 

Collection for conservation ex situ 28 

Storage as seed 28 

Conservation stands 28 

Dissemination of information 30 

Utilization 30 

The need for international action 30 

Concluding remarks 31 

References 31 

Annex 1. Genetic Resources of Forest Trees: Phases and Operations 35 

Annex 2. Genetic Resources of Forest Trees: Duration of necessary phases; 

hypothetical example for a tropical pine 36 

Annex 3. Progress in the Conservation and Utilization of Forest Tree 

Genetic Resources 37 

- 24 - 


Increase in the world's population, together with higher standards of living, result 
in continuous pressure to transfer areas previously under forest to agricultural or other 
uses (Villan 1973). The resulting large-scale disappearance of natural forests is leading 
to an accelerated loss of valuable or potentially valuable germplasm. This loss 5s of 
particular concern in those areas in which botanical and genecological exploration has not 
been systematically carried out, and in which species composition as well as inter- and 
intra-specif ic variation therefore are not known sufficiently well to enable timely and 
adequate conservation measures. 

In addition to the fact that large areas of forests are being partially or fully 
destroyed, areas destined to remain as forest are often being brought under more intensive 
forms of management which may endanger certain species and change the genetic composition 
of others (Kemp et. al. 1976). Even where the central part of the range of a species is 
unnaffected, particular sub-population or provenances at the limits of the species range 
may be in critical danger; it is often these marginal or isolated populations which have 
developed, through natural selection, specific characteristics such as tolerance to drought 
or other adverse environmental conditions, and which may be of great potential value for 
sites with a similar selection pressure. 

The continuous pressure for land mentioned above coupled with an increasing demand 
for wood and wood products has tended to shift the emphasis from utilization of the often 
complex natural forests to plantations of species relatively easy to manage and capable of 
producing large quantities of wood per unit area (Willan 1973). Although the creation of 
plantations to a certain extent will relieve the pressure on the natural forests and the 
genetic material they contain, their establishment may raise problems. The use of plantations 
gives the forester an opportunity to exercise much stricter control not only over" site 
characteristics but also over the genetic quality of his forests (Willan and Palmberg 1974). 
This leads to a development from the use of 'wild 1 to the use of more 'advanced 1 populations, 
in which gene frequencies have been changed to meet specific requirements. In these new 
populations, selected and bred for uniformity, high yield and other short-term objectives, 
the genetic base is often narrowed down to very low levels by restricting the genepool from 
which parental material is drawn and by subsequently rejecting, through selection in given 
conditions, a great proportion of the original population. While adaptation of these new 
populations to specific plantation conditions is increased, their genetic flexibility and 
their potential for future adaptive change to meet often unforseen or unforseeable environ- 
mental changes such as a shift in the average site quality of plantations, the emergence of 
new or genetically adapted pests and diseases, or increased levels of industrial pollution, 
is gradually decreased. The narrowing of the genetic base in populations used for the 
production of seed for future plantations does not, as such, necessarily have negative effects, 
as long as the genetic diversity of the species and the provenances is safeguarded through 
conservation measures in situ or through the establishment of genetic reserves, conservation 
stands and/or widely-based breeding or base populations, from which material to meet new 
requirements can be drawn. 


Conceptually, the issues of genetic conservation are similar whether dealing with 
annuals or long-lived trees, domesticates or wild plants; needs, strategies and methods 
differ in detail, but not in principle (Frankel 1978). 

The exact strategy of conservation depends on the nature of the material and the 
objective and scope of conservation. The nature of the material is defined by the length of 
the life cycle, the mode of reproduction, and the ecological status of the individuals (wild, 
domesticated); the objective could be research, static or evolutionary conservation (see 
below), selection and breeding, etc.; the scope refers to timescale and area considered. 
(Frankel 1970). 

Activities commonly recognized as essential steps to maintainance of genetic diversity 
within individual species and to the fuller use of existing genetic resources are (i) explo- 
ration; (ii) collection; (iii) evaluation; (iv) conservation and (v) utilization. (FAO, 
I975a - See Annexes 1 and 2). 

- 25 - 


Efficient use of existing genetic resources can only be achieved if sufficient infor- 
mation is available on their extent, structure and composition (Brazier <et. aU 1976; Sneep & 
Hendrikson 1979; Lamprey 1975). For a large number of tree species, especially for species 
growing in the tropics, there is a great lack of knowledge on the ecology and biology, as well 
as on their potential as plantation species and the potential use of non-wood products derived 
from them. Even for species of proven value, their variation throughout their natural range 
has often been explored inadequately. This lack of information is well illustrated by the 
fact that every one of the tropical/sub-tropical forest tree species explored and collected 
during the present decade in a programme coordinated by FAO has, as a consequence of the 
exploration, been found to be in danger of depletion, extinction or contamination of its 
genetic resources in at least parts of its natural range; where the genepool is not thought 
to be in danger of extinction, the population has often been so heavily reduced that supplies 
of seed are very limited and may deplete further in future (Keiding and Kemp 1978). 

For practical purposes, the field activities in the fundamental step of exploration 
can be divided into (i) botanical and (ii) genecological exploration. Botanical exploration 
includes the correct taxonomic identification of species and knowledge of the limits of their 
distribution, with particular reference to isolated occurrances. For some forest tree 
species adequate information has been available well before the start of genecological ex- 
ploration, for others it may be necessary to combine the two operations together. Botanical 
exploration logically leads to species trials. 

Through genecological exploration patterns of ecological and phenotypic variation 
within the natural range of species are studied, leading to provenance seed collections 
and provenance evaluation (FAO 1975). 


Collection for evaluation consists in collecting relatively small samples of seed from 
a relatively large number of seed sources, covering the whole natural range of the species. 
In the initial stage, collections thus comprise range-wide sampling on a fairly coarse grid; 
in some cases a second stage, sampling a limited part of the range on a finer network, may 
be called for after the results of first stage provenance trials are available. In these 
second-stage collections seed is often kept separate by mother trees to enable the evaluation 
of genetic variation within, as well as between, provenances. The genepool s included in 
the collections may be indigenous, or introduced. In forestry, these latter so-called 'land 
races 1 (i.e. exotic plantations which to various degrees have adapted to local conditions 
as a response to natural and sometimes artificial selection), are potentially of great 
importance as sources of seed, and should be included in the collections (FAO 1975a, 
Turnbull 1978. 

To determine the number and location of populations to be sampled, environmental 
gradients are generally followed; sampling within each population could be done at random 
or selectively. Although the latter option is often used in within population sampling 
it should be remembered that phenotypic superiority is no guarantee of genetic superiority, 
specially in cases in which the past history of the population is unknown (Barner 1974; 
Bennett 1970). 


Collection of range-wide samples should be followed by the establishment of provenance 
trials aimed at revealing potentially useful variability, degree of adaptation to a range 
of environmental conditions, and economic or social value of the species/provenances under 
test. The evaluation should be carried out on as many potential sites as possible and, 
whenever feasible, be centrally coordinated. 

- 26 - 

The development of the concept of genetic conservation in the 1950s was mainly due 
to the realization that the primitive cultivars of traditional agriculture were rapidly 
disappearing, and the genetic diversity accumulated in them over many centuries substituted 
by varieties selected and bred to meet short-term needs. The importance of maintaining 
genetic diversity and genepools from which new genes can be introduced into existing 
varieties in order to improve adaptation, yield and resistance to diseases and adverse 
conditions, has clearly been brought into focus by some major outbreaks of diseases, espe- 
cially in food crops (Frankel 1978, Sneep and Hendriksen 1979). Not only should variability 
within species of known value be conserved, but maximum diversity should also be maintained 
on a species level, including as yet unknown and untested material, thus keeping future 
options open (Whit mo re I975a). 

Conservation, used in its proper sense, embraces both preservation and utilization; 
conservation is, in fact, an aspect of resource management which ensures that utilization 
of the resource is sustainable, at the same time safeguarding genetic diversity essential 
for its maintenance. 

A compromise between biological, technical, economic and administrative factors is 
often unavoidable when choosing long-term strategies for the conservation and utilization 
of genetic resources. The final objective must be to choose methods which minimize losses 
and maximize gains in terms of usefulness, knowledge and integrity (Frankel 1970a), 

Current problems in genetic conservation are often so severe that it is tempting 
to concentrate on them alone. However, strategies for action should include preventive 
measures through the inclusion of conservation aspect in long-term planning at the policy- 
making, the organizational as well as the technical level (Anon 1980). 

The main strategies for conservation have been outlined as follows (Burley and 
Styles 1976): 

1. Conservation of ecosystems. Such conservation of carefully selected areas of 
adequate size and with suitable management policies would preserve not only 
forest trees but also other elements of the ecosystem (plants, mammals, birds, 
etc.), as well as other potentially valuable products such as extrativas, 
fruits, etc. 

2. Preservation of rare species and species threatened with extinction. This 
goal could often be achieved under general ecosystem conservation, if carried 
out competently. 

3. Prevention of "genetic erosion 11 , i.e. depletion of genetic variability. In 
this field it is not enough to conserve a species as such; we must make certain 
to conserve a broad spectrum of genetic variability to act as a reserve for 
present and future needs (adequate seed sources, wide genetic variability as 

a base for tree breeding , etc.). Such material can be conserved in in s i t u 
reserves, or can be sampled by means of seed, pollen or vegetative material, 
in such a way that the conservation of the major part of the genetic variability 
is insured. Seed, pollen or other material can either be stored as such, or 
be used for the establishment of conservation stands ex situ. 


Conservation jln situ, i.e. conservation of species /provenances as part of a viable, 
existing ecosystem, is generally the most desirable method of conserving forest genetic 
resources, provided that the area can be given full protection and provided the genetic 
material conserved is made available for use both within and outside the country of origin 
(FAO 1975; Whitmore 1975a, b; Lamprey 1975; IUCN 1978). For many species, e.g. for a vast 
number of rainforest species which are not fast-growing pioneers, which occur as individuals 
rather than stands, and for which knowledge on ecology and genetics is scarce or lacking, 
in situ is the only method of conservation available to us at our present level of 
knowledge (Kempt 1978). 

- 27 - 

In situ conservation of forest genetic resources will often for practical reasons 
have to be combined with other environmental, scientific or socio-economic purposes, this 
generally means that a compromise must be made between various objectives of the reserve. 

Genepool conservation frequently deals with genetic differences which cannot be 
directly identified but only surmised. It is concerned with population samples, possibly' 
along latitudinal or altitudinal transects, often over extensive areas, to include a spec- 
trum of ecological variability to provide a corresponding spectrum of genetic variability. 
The efficiency of ecosystem conservation (e.g. biosphere reserves, national parks) to 
adequately meet the needs of genepool conservation is closely related to the size, number, 
distribution and location of these reserves. 

There is a general consensus that the conservation of representative samples of 
most ecosystems will require an area within the range of 100 to 1 000 ha, the exact size 
depending on the heterogenity of the area as well as on its species composition (Ashton 
1976). If conservation of genetic resources is a major objective of ecosystem conserva- 
tion, the inclusion within the reserve of the minimum number of interbreeding individuals 
needed for a viable genepool (i.e. a population which is able to retain its self-renewing 
capacity), must be considered rather than the total area of the reserve per s (Roche 1975), 

Considering genetic resources at a species level , Ashton (1976) working with rain- 
forest species in Borneo made a theoretical estimation of the area of forest needed for 
conservation, arbitrarily assuming that 200 mature individuals will form a viable popula- 
tion; by this criterion an area of at least 2 000 ha of unmodified virgin forest would 
have been needed to conserve the tree species of two areas examined, while only 60% of the 
species would have been safeguarded in 1 000 ha. At the intra-specific level Dyson (1974), 
studying numbers of individuals quoted for effective breeding populations in animals, 
estimates that 200 individuals will constitute a minimum "safe" population for forest 
trees for maintaining genetic variability, provided that sampling is done from at least 
three parts of the range. Up to 25 000 individuals are recommended by Marshal (quoted 
in Kemp 1978) 'for maintaining a given level of heterozygosity in populations in "a dif- 
fusely outbreeding tree species'. However, as discussed below, the conservation of a 
mythic heterozygosity (Namkoong 1979a) by conserving specific genotypes is neither 
desirable nor possible. 

Theories as to the relative merits of one large as opposed to several smaller, 
separate reserves, have been extensively discussed. The answer will depend on the exact 
objectives of conservation, the amount of inter- and intra-specific variation to be con- 
sidered, and the distribution of gene frequencies. From a purely managerial point of 
view, one or very few large areas would be preferable, as a large number of scattered re- 
serves are difficult to manage and protect. However, especially in the case of areas with 
a complex species composition, and in the case of conserving the intra-specific variation 
of widely distributed species, a series of reserves, strategically placed to sample the 
full range of ecological variation is called for. It is particularly important to include 
extreme environments and marginal populations, in which selection effects may have created 
varieties or ecotypes of particular potential value, and in which gene frequencies may 
differ from those in the main population, giving us bigger chances to capture 'rare genes' 
(Namkoong 1979a, b). 


Although conservation In situ in theory may be the most effective strategy it may, 
in reality, face enormous difficulties which often are social, political or financial 
rather than technical in nature (Sastrapradja et al. 1978; Kemp et. al. 1976). The alter- 
native way of conservation is ex situ. Ex situ conservation is especially useful when 
dealing with certain species or genera with a combination of biological characteristics 
which make them amenable to this approach; through knowledge of the breeding system and 
biology of the species, as well as of the methodology for growing them in plantation con- 
dition and/or storing their seed, are prerequisites for using this strategy. Many of the 
species which during recent years have attracted the attention of foresters for their use 
in high-yielding plantations fall into this category. 

Sometimes, especially in the case of economically valuable plantation species ex- 
tensive genetic modification of native stands caused by man may occur (Libby 1978). This 
situation arises if non-native populations of a species are used for seed procurement and 
the subsequent establishment of plantations close to existing native stands. Clouds of 

- 28 - 

pollen from the plantations will repeatedly be dispersed over the native stands, resulting 
in offspring which will be increasingly contaminated by genes from the foreign populations, 
gradually leading to a loss of the original genepool. In these cases, in situ conservation 
will not be feasible and the only way to conserve the original population will be by ex 
situ conservation. 


Where the exploration phase has shown that certain populations are endangered but 
that conservation ^n situ is not feasible, early collection is necessary of substantial 
quantities of seed or other propagating material of endangered provenances either for tem- 
porary storage or for immediate establishment of ex situ conservation stands on new sites 
(FAO 1975). Sampling for variation (i.e. random rather than selective sampling) is 
essential for conserving the integrity of allele frequencies (Frankel 1970b). 

Methods of sampling for gene conservation purposes and theoretical number of indivi- 
duals needed to maintain intraspecific allelic variation are discussed by Namkoong (1979a), 
who gives probability levels for losing specific alleles occurring at given frequencies, 
using different sampling intensities within and among populations. 

It is not possible to lay down general rules and guidelines for sampling, as many 
factors, both interrelated and independent, affect the intraspecific variation we are 
trying to capture (heterogenity and size of the natural range, ecology, breeding system 
and population structure of the species, etc.). However, as no system of sampling and 
collection as likely to achieve the objective of saving all allelic combinations present 
in a species, sampling is generally aimed at saving as many of the existing alleles as 
possible for future recombination and use (Namkoong I979a). We should thus aim at conserving 
and evaluating genes rather than genotypes. 


In addition to being a means of conservation per se, storage of seed is often an 
essential link between collection and later field operations. Meticulous handling of the 
seed during all phases of the work is essential. For many species particularly in the 
tropics, there is inadequate knowledge of practical methods of both long-term storage, and 
additional research is urgently needed. 

Conservation of forest trees is generally done as conservation stands jln situ or 
x situ, rather than as seed, as is often the case in agricultural crop species. This 
difference in approach is mainly due to practical difficulties: plant gene banks should 
regenerate their seed collections whenever the viability falls by, at most, fifteen percent 
below the initial value at which the seed was stored (Wang 1976; IBPGR 1976). With the long 
vegetative period that most tree species have before they produce viable seed, seed regenera- 
tion by growing a crop and re-collecting will be a prolonged and expensive procedure; in 
addition, natural selection during this prolonged period is likely to have greater effects 
on genetic composition than in the case of species which produce seed within a short period 
of time after sowing. 

With present-day knowledge of the physiology and biochemistry of pollen and tissue 
culture, conservation of forest genetic resources in these forms seems unlikely to become 
more than a useful supplement to other forms of conservation. Although pollen storage 
is a valuable means for short or intermediate-term conservation storage time of pollen, 
when using known drying and storing techniques, is generally shorter and less reliable than 
that of seed. Likewise, and with the possible exception of vegetatively propagated species* 
the conservation of forest genetic resources by means of tissue culture is generally not 
considered likely to become of great importance (Wang 1978; Frankel 1978). 


Conservation stands ejc situ are expensive to establish and to maintain, and therefore 
they are normally confined to species of proven or evident potential value (FAO 1975). 
Danger of extinct ion, economic potential and difficulty in reprocurement of seed should be the 
main criteria in forming priority list of species and provenances for jnc situ tonservation.- 


- 29 - 

Guldager (1978) lists four conservation objectives which may be met by the establishment 
of ex situ conservation stands: 

(1) Static conservation, in which the genotype frequencies of the original 
population are maintained. As discussed above, this method is not practicable for most 
forest tree species. 

(2) Static conservation, in which the gene (allele) frequencies of the original 
population are maintained. No genetic information is lost and any genotype found in the 
original population could in principle be reproduced although genotype frequencies in the 
ex situ stands might be different than in the original population. 

(3) Evolutionary conservation, in which gene frequencies in the stand are allowed 
to change according to natural selection pressures. 

(4) Selective conservation, in which gene frequencies in the stand are deliberately 
changed by man in order to capture characteristics important for plantation economy in a 
region, and at the same time eliminate undesirable characteristics. To avoid a decrease 

of the genetic potential for future plantation establishment in environments different from 
the environment in which this type of stand is established, replication is necessary in 
each potential plantation area. In the long term, selective conservation faces the same 
problem as met in long term breeding programmes, (i.e. maintenance of genetic variation, 
avoidance of inbreeding). 

The ex situ conservation stands known to have been established to date fall under 
categories IT) and (4). The level of maintenance of genetic integrity in these stands will 
be dependent on three main factors (Guldager 1976): (i) sampling in the original popula- 
tion; (11) survival and growth ex situ of sampled genotypes (i.e. adaptation to new selection 
pressures); (iii) mating patterns ejc situ between the sampled genotypes. 

(i) Sampling for conservation has been discussed above* Even the most rigorous 
efforts to maintain original gene frequencies through careful management in a series of 
conservation stands ex situ will be of little avail from the point of view of species/ 
provenance conservation if the gene frequencies have already been considerably changed 
during initial sampling. Sampling is thus of critical importance. Long term storage or 
maltreatment of seed are other factors that may critically influence gene frequencies even 
before the stands are established. 

(ii) For most plantation species it is possible to combine the selection of suitable 
sites with efficient nursery and plantation techniques so as to ensure almost 100% survival 
in the field. Initial competition between genotypes can be minimized by wide spacing. The 
question of mechanical versus silvicultural thinning in the stands will depend both on the 
ultimate objective of conservation and on practical possibilities. However, if the stands 
are established over a wide range of sites in which environmental pressures vary, a great 
proportion of the genetic variation is likely to be maintained even if thinning is carried 
out in favour of desirable phenotypes. As a compromise, a proportion of the trees to be 
left (say, 1%) may be phenotypically selected before carrying out a systematic thinning 
for the rest of the stand. This approach will be adapted in the case of the international 
conservation stands mentioned below. 

(iii) Our possibilities to accurately pass on genetic information from the first 
generation ex situ conservation stand to the next depends on mating within the stand 
(synchronization of flowering, proportion of actual random mating, etc.), size of population 
(influencing genetic drift and inbreeding coefficient), and migration in terms of pollen 
contamination. To help overcome these problems, location (near optimum or optimum for 
flowering and seed production), size (recommended size is 10-30 ha FAO 1975, 1977) and 
isolation (300 m or more between hybridizing species/provenances; FAO 1977), should be 
carefully considered. In addition to careful siting, the conservation stands require 
meticulous standards of site preparation, planting and tending (FAO 1975). An imperative 
condition for their establishment in a region is thus that sufficient technical expertise 
as well as organizational stability is available, to ensure a high standard of long tern 
management. Interest in the provenances concerned from a plantation point of view, is 
likely to increase the benefit as well as the security of the scheme (Guldager 1978). 

- 30 - 

Detailed recommended prescriptions for the establishment and management of ex situ 
conservation stands have been published in Annex 7 of the Report of the 4th Session of the 
FAO Panel of Experts on Forest Gene Resources (FAO 1977). 

Besides the long-term benefits of conserving species/provenances of known genetic 
characteristics, conservation stands have valuable possibilities for short-term utilization, 
providing seed and other genetic material for immediate use. In cases where international 
finance has been provided, agreements have therefore been drawn up to ensure that the stands 
benefit all countries interested in the species/provenances (see FAO 1977, Appendix 7). 


Another aspect of conservation is the preservation and dissemination of information. 
It is important to preserve not only the areas, units, populations or individuals, but also 
the relevant data on them. This data must be carefully recorded, safeguarded and dissemi- 
nated (Frankel 1970a). 


Utilization is the ultimate objective of all activities concerned with forest genetic 
resources. It comprises both the use of bulk supplies of seed or other propagating material 
for large-scale plantation schemes, and the breeding of better adapted and more desirable 

As information becomes available from provenance trials as to the most suitable seed 
sources, emphasis will gradually switch from evaluation to the utilization of bulk supplies 
of seed or sometimes other propagating material of populations found well-adapted to given 
conditions. The supply of bulk quantities of propagating material should be the* responsibi- 
lity of Government Forest Services or commercial seed merchants, although agreements both 
within and between countries on common standards of genetic and physiological quality of 
the material are essential (FAO, 1975a). 

Individual selection and breeding within locally adapted provenances provide a method 
of achieving additional improvement in selected characteristics. In the case of introduced 
species, an important interim stage between successful provenance trials and large-scale 
afforestation with the best-adapted provenances may be the establishment of one or more 
substantial blocks (5 ha or more) of these provenances to act as seed stands and as a basis 
for local selection and breeding* The same stands may combine the purpose of ex situ 


When the urgency of conservation and the massive efforts needed are jointly con- 
sidered, it becomes evident that conservation of the world's genetic resources requires the 
cooperation of all nations. 

Although progress in development of improved forest genetic resources will remain 
largely dependent on the active efforts of individual countries or research institutes, 
these can only be fully effective in an international framework (FAO 1975). Maintenance 
of the genetic diversity of species either jln situ or eoc situ may have to be spread over 
many environments in a number of countries, collection of seed cannot be confined within 
national borders; research coordinated to give information of species/provenance performance 
over the maximum range of sites possible will be of great mutual benefit to cooperating 
institutes and nations; safety and permanence for irreplaceable collections of genetic 
material, either j.n situ or ex situ should be secured for perpetuity by agreements under 
international supervision. 

Many countries which contain forest genetic resources of great but sometimes un- 
explored potential value are at an early stage in their economic development. There is 
often a severe shortage of funds and skilled staff in the forestry sector and those 
available are, logically, channeled to meet immediate national needs when e.g. drawing up 
species priorities. It is therefore highly desirable that international resources should 
be made available to help in the development of strategies and to safeguard material 
invaluable to many nations. 

- 31 - 

There is another aspect to conservation, and that is the conservation and dissemina- 
tion of information. It is not only important to conserve areas, communities, populations 
and individuals, it is also important that information related to them is adequately re- 
corded, safeguarded and made available (Frankel 1970). 

The best way of ensuring efficient coordination in the wide field of forest genetic 
resources is by the adoption of a global programme such as that proposed by the FAO Panel 
of Experts on Forest Gene Resources (FAO 1975a). Such a program should ensure the inte- 
gration of conservation measures with the equally important activities of exploration, 
collection and utilization. At the same time it should improve efficiency through coor- 
dinating the efforts not only of the many countries but also of the several international 
agencies concerned with genetic resources (Roche 1978). 

Progress in the conservation of forest gene resources in the last ten years is out- 
lined in Annex 3. 


In the rapidly advancing field of forest genetics we have answered many question 
in recent years, but these answers have often led to new and more difficult questions. 
We have learned enough techniques to be certain that we can develop new breeds to more 
accurately meet present-day needs. We have also come to realize more fully that the ori- 
ginal gene pools will be lost unless positive action is taken to conserve them. Now we need 
to decide how to organize breeding and gene management strategies that will meet both 
immediate needs and long-term requirements (Namkoong 1978). 

It should not be difficult to conduct tree improvement programmes which include 
short-term and long-term objectives side by side, provided that those responsible for 
planning and finance realize that the long-term programme is no less important and no less 
deserving of funds than the short-term one, and that the greater the genetic diversity we 
can maintain and save now, the wider will be our options for finding suitable genetypes 
to meet future needs (Namkoong 1979). 


Anon, 1980. World Conservation Strategy. Living Resource Conservation for Development E, 
F, S. IUCN/UNEP/WWP/FAO (en prensa). 

Ashton, P.S. 1976. Factors Affecting the Development and Conservation of Tree Genetic 
Resources in South-East Asia. En; Tropical Trees: Variation, Breeding and Conservation 
(Eds. J. Burley & B.T. Styles). Linnean Society, Oxford, U.K. 

Barner, H. 1974. Classification of Sources for Procurement of Forest Reproductive Material. 
En; Report of the FAO/DANIDA Training Course on Forest Tree Improvement. Limuru, Kenya, 
September-October 1973. FAO/DENTF 112. FAO, Roma. 

Bennett, E. 1970. Tactics of Plant Exploration. En; Genetic Resources in Plants- their 
Exploration and Conservation (Eds. O.H. Frankel & E. Bennett). IBP Handbook No. 11. 
Blackwell Scientific Publications. Oxford and Edinburgh. 

Brazier, J.D., Hughes, J.F. & Tabb, C.B. 1976. Exploration of Natural Tropical Forest 
Resources and the Need for Genetic and Ecological Conservation. En: Tropical Trees: 
Variation, Breeding and Conservation (Eds. J. Burley & B.T. StylesT* Linnean Society, 
Oxford, U.K. 

Burley, J. & Nikles, D.G. 1972 & 1973a. Selection and Breeding to Improve some Tropical 
Conifers. Vols I & II. Based on Papers submitted to a Symposium organized by IUFRO 
Working Parties S2.02.08 and S2.03.01, held in Gainesville, Florida, USA in 1971. 
Commonwealth Forestry Institute, Oxford, U.K. 

Burley, J. & Nikles, D.G. I973b. Tropical Provenance and Progeny Research and International 
Cooperation. Based on Papers submitted to a Symposium organized by IUFRO Working Parties 
S2.02.08 and S2.03.01, held in Nairobi, Kenya in 1973. Commonwealth Forestry Institute, 
Oxford, U.K. 

- 32 - 

Bur Icy, J. & Styles, B.T. (Eds.) 1976. Tropical Trees: Variation, Breeding and 
Conservation. Linnean Society, Oxford. 

Cromer, D.A.N. 1976. Report on Consultant Mission on Conservation of Forest Genetic 
Resources in selected countries in Asia. FO:MISC/76/27. FAO, Rome. 

Dyson, W.G. 1975. A Note on the Conservation of tree species in situ. In: Report of the 
Third Session of the FAO Panel of Experts on Forest Gene Resources. FO-FGR/3/Rep. FAO 

FAO, 1973-79. Forest Genetic Resources Information. Forestry occasional Paper. FAO Rome. 

FAO, 1969. Report on the First Session of the FAO Panel of Experts on Forest Gene 
Resources. FO:FGR/l/Rep. FAO, Rome 

FAO, 1972. Report on the First Session of the FAO Panel of Experts on Forest Gene 
Resources. FO:FGR/2/Rep. FAO, Rome. 

FAO, 1975a. Proposals for a Global Programme for the improved utilization of forest 
genetic resources. Forest Genetic Resources Information No. 4 1975/1 FAO, Rome. 

FAO, 1975b. Report on the First Session of the FAO Panel of Experts on Forest Gene 
Resources. FO:FGR/3/Rep. FAO, Rome 

FAO. 1977. Report on the First Session of the FAO Panel of Experts on Forest Gene 
Resources. FO:FGR/4/Rep. FAO, Rome 

Frankel, O.H. 1970b. Evaluation and Utilization - Introductory Remarks. In: Genetic 
Resources in Plants - their Exploration and Conservation (Eds. O.H. Frankel and E. 
Bennet). IBP Handbook no. 11. Blackwell Scientific Publications, Oxford and Edinburgh. 

Frankel, O.H. 1970a. Genetic Conservation in Perspective. In; Genetic Resources in 
Plants - their Exploration and Conservation (Eds. O.H. Frankel and E. Bennett). IBP 
Handbook no. 11. Blackwell Scientific Publications, Oxford and Edinburgh. 

Frankel, O.H. 1978. Philosophy and Strategy of Genetic Conservation in Plants. Tercera 
Consulta Mundial sobre el Mejoramiento de Arboles Forestales. FC:FTB/77-l/2. Canberra, 

Franke, C.H. & Bennett, E. (Editors) 1970. Genetic Resources in Plants: their Exploration 
and Conservation. IBP Handbook no. 11. Blackwell Scientific Publications, Oxford and 

Guldager, P. 1978. Ex situ Conservation stands in the tropics. In; The Methodology of 
Conservation of Forest Genetic Resources. FO:MISC/75/8. FAO. Rome. 

IBPGR, 1976. Report of IBPGR Working Group on Engineering, Design and Cost Aspects of Long- 
Term Seed Storage Facilities. International Board for Plant Genetic Resources. FAO, Rome. 

IUCN, 1978. Categories, objectives and criteria for protected areas. A Final Report 
prepared by Committee on Criteria and Nomenclature Commission on National Parks and 
Protected Areas. Morges, Switzerland. 

Keiding, H. and Kemp, R.H. 1978. Exploration, collection and investigation of gene 
resources: tropical pines and teak. Tercera Consulta Mundial sobre el Mejoramiento de 
Arboles Forestales. FO-FTB-77-1/3. Canberra, Australia 

Kemp, R.H. , 1976. Report on Consultant Mission on Conservation of Forest Genetic 
Resources in selected countries in Africa. FO:MISC/76/26. FAO, Rome. 

Kemp, R.H., 1978. Exploration, utilization and conservation of genetic resources. Third 
World Consultation on Forest Tree Breeding. FO-FTB.77.1/1. 

- 33 - 

Kemp, R.H. , Roche, L. & Will an, R.L., 1976. Current activities and problems in the 
exploration and conservation of tropical forest gene resources. In: Tropical Trees: 
Variation, Breeding and Conservation (Ed. J. Bur ley & B.T. StylesTT Linnean Society, 
Oxford, U.K. 

Kemp, R.H. & Whltmore, T.C., 1978. International Cooperation for the Conservation of 
Tropical and Sub-Tropical Forest Genetic Resources Exemplified by South East Asia. 
Eighth World Forestry Congress. FQL/26-11. 

Libby, W.J., Krafton, D. & Fins, L., 1978. California Conifers. In; The Methodology 
of Conservation of Forest Genetic Resources. (Ed. L. Roche). FO:MISC/75/8, FAO/UNEP, 

Lamprey, H.F., 1975. The distribution of protected areas in relation to the needs of biotlc 
community conservation in Eastern Africa. IUCN Occasional Paper no. 16. Merges, 

Namkoong, G., 1978. Selecting Strategies for the Future. Third World Consultation on 
Forest Tree Breeding. FO-FTB-7 7-6/1. Canberra, Australia. 

Namkoong, G. , 1979a. Methods of pollen sampling for gene conservation. Chapter 17. 
Pollen Management Handbook. Southern Forest Tree Improvement Committee (en prensa). 

Namkoong, G, 1979b. Introduction to quantitative Genetics in Forestry. USDA, Forest 
Service Technical Bulletin No 1588. Washington D.C. 

Nikles, D.G., Burley, J.& Barnes, R.D. (Eds), 1978. Progress and Problems of Genetic 
Improvement of Tropical Forest Trees. Proceedings of a Joint Workshop of IUFRO Working 
Parties S2.02.08 and S2.03.01, held in Brisbane, Australia, 4-7 April 1977. Vols. I and 
II. Commonwealth Forestry Institute, Oxford, U.K. 

Roche, L. (Editor), 1978. The Methodology of Conservation of Forest Genetic Resources. 
Report on a Pilot Study. FO:MISC/75/8. FAO/UNEP, Roma. 

Sastrapradja, S. ejt al, 1978. The conservation of forest animal and plant genetic 
resources. Eighth World Forestry Congress. FQL/26-0. 

Sneep, J.& Hendriksen, A.J.T. (Eds), 1979. Plant Breeding Prospectives. Centre for 
Agricultural Publishing and Documentation, Wageningen, Netherlands. 

Turnbull, J.W., 1978. Exploration and Conservation of Eucalypt Gene Resources. Tercera 
Consulta Mundial sobre el Mejoramiento de Arboles Forestales. FO-FTB-7 7-1 /A, Canberra, 

Wang, B.S.P., 1978. Tree Seed and pollen storage for genetic conservation: possibilities 
and limitations. In; Methodology of Conservation of Forest Genetic Resources. (Ed. 
L. Roche). FO:MISC/75/8. FAO/UNEP, Rome. 

Whltmore, T.C., 1975a. Conservation review of tropical rain forests: General Considera- 
tions and Asia. IUCN, Morges, Switzerland. 

Whitmore, T.C., 1975b. Tropical Rain Forests of the Far East. Oxford University Press, 

Wlllan, R.L., 1973 Forestry: Improving the Use of Genetic Resources. Span 16 (3):119-122, 

Willan, R.L. & Palmberg, C., 1974. Better Use of Forest Genetic Resources. En: Report 
on the FAO/DANIDA Training Course on Forest tree Improvement. Kenya, 1973. 
FAO/DEN/TF 112. FAO, Rome. 

of Finns Carbaea var . liondmrensl s 

C CON ARE, diaguairainas) 

- 35 - 


Phases and Operations I/ 


V Willan and 

- 36 - 







^ i 






























p. c 


. j 




























,._ J 








CO fH 

rvj rv 

S 6 





QE a 

5 M 


L^ ' 










S *""! 



^H ^A C 





i & 



i ** 





E ** 


2 i 


a -c 


S pL 
















O Q 

































i ^ 




h 1 



4* 1 




(Tear - beginning 

oologioal explorat 

*^_ , ^ 

- Botanical 
(0-200 years) 



( *1 

l j 




f 1 




3 c 

1 H 











J | 


vw** na wi*M0n v A A, *w 

- Seed* pollen 






urn HO* Houaanoo 




D 1 


5 < 








U Selection of plus 

1 4*i i_.t..._1 -4. 













- 37 - 

Annex 3 


Concern over the loss of genetic diversity has grown rapidly since early fifties, 
and has spurred increasing national and international action. 

Some national institutes had already begun systematic forest tree seed collections 
for international use at that time, but the present picture has spurred international and 
national exploration, collection, conservation and research efforts in forest tree genetic 
resources . These efforts have created world awareness of the need to conserve these 
resources, mobilizing national and international funds for experimental plans, pilot studies 
and practical activities in this field. 

The panel has met four times. FAO has published information on these meetings, 
indicating progress, past and present trends, and recommendations for future action (FAO, 
1969, 1972, 1975b and 1977). So far, the funds recommended by the panel for programmes 
coordinated by the FAO Forestry Department have been mainly earmarked for the exploration 
and collection phases to fund certain organizations working in these fields. FAO cooperates 
not only with national institutes, but with other international agencies such as Unesco , 
IUCN - , and UNEP $.' , and collaborates actively with the competent IUFRO - working parties. 
Recently some funds have been received from the International Board of Plant Genetic 
Resources (IBPGR), an auxiliary body of the Consultancy Group on International Agricultural 
Research (CGIAR), for the purpose of mobilizing long-term financial support to close the 
gaps in agricultural research in the developing countries. 

Through the Panel, priorities have been established by region and by species for 
each phase of a programme on forest tree genetic resources (FAO, 1977, Appendix 8). 
These priorities, which are periodically reviewed in the light of the most recent discove- 
ries and measures adopted, are based on the extent to which genetic resources of species 
are endangered, and their potential or actual socio-economic importance. However, as it 
is only in the course of exploration that exact information will be obtained on the 
conservation status of a species, the priorities indicated and species included in the list 
reflect to a certain extent the quantity and quality of the information available to the 
Panel to adopt its decisions, and not only the real situation (Keiding and Kemp, 1978). 

Based on the orders of priority indicated by the Panel so far, full-spectrum ex- 
ploration and collection activities have been carried out, followed by the establishment 
of centrally coordinated international provenance trials for 12 tropical species. Conside- 
rable progress has also been made in the exploration, collection, distribution and evalua- 
tion of various genera, including Tectona, Populus, Pinus, Pseudotsuga, Araucaria, and 
Eucalyptus. Among the genera most recently included in the Programme are Acacia, Prosopis, 
Terminalia and Aucoumea. FAO has published abstracts of the most important collections 
(FAO 1975b; FAO 1977) . The reader may also consult the FAO review "Information of Forest 
Tree Genetic Resources". 

Although many of the species trials based on range-wide collections carried out in 
recent years are too new to supply precise data, many already indicate the existence of 
major provenance differences and clear interactions between provenance and environment, 
confirming that research on provenance is as important for tropical species as it is for 
temperate species. The findings of many of these trials have been published in summary 
form by species and countries in the proceedings of the joint IUFRO Working Parties S2.02.08 
and S2.03.02, held in 1971, 1973 and 1977 (Bur ley and Nikles, 1972, 1973a, 1973b; Nikles, 
Bur ley and Barnes, 1978). 

I/ United Nations Educational, Scientific and Cultural Organization. 

2/ International Union for the Conservation of Nature and Natural Resources. 

3/ United Nations Environment Programme. 

4/ International Union of Forestry Research Organizations. 












X 0. 


o o 











- 39 - 

In 1975, with financial support from UNEP, FAO carried out a pilot study which 
gave rise to the publication "Methodology of the Conservation of Forest Tree Genetic 
Resources" (Roche, 1978a). Based on the technical recommendations of this study and on 
the early findings of the exploration and international provenance trials mentioned above, 
in 1975/76 an FAO/UNEP project was launched for the conservation of genetic resources of 
certain species and provenance. This experimental project, which includes elements of 
both on-site and off-site conservation, is now drawing to a close, and a report on its 
achievements and conclusions is now being prepared* 

The off-site component of the above FAO/UNEP project has performed as expected. 
Over the last four years, 33 international ex situ conservation/selection stands, each 
10 ha in size, have been established in five African countries and one Asian country, with 
a total of 11 provenances of four different species (see Table 1). In addition to the 
international stand funded by UNEP and FAO, many of the countries participating in the 
project, as well as some neighbouring countries, have established national ex situ selec- 
tion/conservation stands. In Appendices 7/1 to 7/4 of the report of the Fourth Session of 
FAO Panel of Experts on Forest Genetic Resources (FAO, 1977) include the reasons for the 
selection of the species /provenance for the project, the agreement between FAO and the par- 
ticipating countries, some recommendations on the establishment, management and handling 
of stands, and cost estimates. Succinctly, the agreement calls for international finan- 
cing to foot the bill for seed costs, plus the projected cost of the first two years of 
the establishment phase. The host government pledges to supervise the establishment, main- 
tenance and management of the stands, and to make 50 percent of the seeds, or other repro- 
ductive material collected, available to other countries at cost price. 

DANIDA (Danish International Development Agency) began a complementary project in 
1979 on off-site conservation. 

It has proven difficult to reach agreements on in situ conservation. The FAO/UNEP 
project only provided funds for two botanical reserves In Zambia for the rn situ 'conserva- 
tion of Baikiaea plurijuga (Zambia Sequoia) or Zambesi Redwood. The main reasons why 
identifying proper zones for on-site conservation is a problem are listed below: 

1) On-site conservation in the tropics generally has to contend with heterogeneous 
ecosystems, of which economically valuable species are only a small part. When 
funds are scarce, national priorities for expenditures and efforts tend to 
centre on other sectors and on other species; 

2) Tropical ecosystems are complicated and not enough is known about them, unlike 
ex situ conservation stands (even-maturing monocrops). They are also difficult 
to manage; 

3) It is difficult to predict when the first substantial benefits from in situ 
conservation will be produced at the national level; 

A) It is often difficult to identify which specific vital operations or phases of 
in situ conservation might be suitable for international, short-term financing; 

5) The establishment of ex situ conservation stands is a specialized type of re- 
forestation, and, as such, is clearly within the sphere of competence of the 
forestry service of the country involved. However conservation of the ecosystem 
(and therefore in situ conservation) is often under some other authority such 
as departments responsible for flora and fauna, national parks, etc. 

There has been no let-up in action to overcome these difficulties. 

Among the progress achieved in the dissemination of information on forest tree 
genetic resources in recent years we may list such meetings as: (1) the three FAO/IUFRO 
World Consultations on Forest Tree Improvement (Stockholm, 1963; Washington, 1969; Canberra, 
1977), which viewed existing data on the scientific principles of forest tree improvement 
and forest genetics; practical advantages and progress in tree improvement; and problems 
and prospects for the utilization and conservation of forest tree genetic resources; (11) 
the three meetings of the IUFRO working parties S2.02.08 and S2.03.01 already mentioned 
in this paper; (ill) the Eighth World Forestry Congress (Indonesia, 1978), which acknowledged 
the fundamental importance of genetic conservation, devoting a session to this item on Its 
agenda. A series of training courses have also been held on tree improvement. 

- 40 - 

These were financed by UNDP and DANIDA and organized and implemented by the FAO Forestry 
Department (Denmark, 1966; U/S/A/, 1969; Hungary, 1971; Kenya, 1973; Thailand, 1975). A 
training course on tree improvement organized by the Forest Research Division of the 
Commonwealth Scientific and Industrial Research Organization, Canberra, and financed by 
the Australian Government, was held in Australia in 1977. The International Seed Trials 
Association, ISTA, has organized various seminars on the processing and analysis of 
forest seeds. 

The FAO publication "Information on Forest Tree Genetic Resources' 1 (FAO 1973-79) 
launched in 1973 publishes three Issues in the biennium. The review features periodic 
reports on these and other meetings, seed collections, provenances and the exploration, 
evaluation, utilization and conservation of forest tree genetic resources , 

- 41 - 


C. Palmberg 

Forest Resources Division 

Forestry Department 




Introduction 41 

Collection for evaluation 41 

Sampling at the population level 42 

Sampling at the individual level 43 

Collection for conservation 43 

Collection for utilization 43 

Sampling at the population level 43 

Sampling at the individual level ... 44 

General Remarks 44 

Bibliography 44 


In any seed collection, sampling will ultimately determine the proportion of the 
genetic variation present that we will capture. Errors or carelessness in sampling* i.e. 
in the selection of the populations and the trees from which seed is collected, cannot 
be remedied at the planting stage, no matter how careful and sophisticated our experimen- 
tal design is or however carefully we establish our plantations. 

In broad terms, sampling is done at two levels, the population level (provenance 
and stand), and the individual level. The choice of the exact method and intensity of 
sampling will depend on the specific objectives of the seed collection. 

The main objectives of seed collection are: 
(i) Evaluation 
(ii) Conservation 
(iii) Utilization (large-scale afforestation). 


Evaluation, in this context, generally refers to the establishment of species and 
provenance trials in which the range and type of variation is assessed and the adaptation 
of the various species/provenances to potential plantation sites is evaluated. 

The two main questions to be answered when planning a specific seed collection 
mission is how to allocate, in practice, the available time and funds between the frequency 
(how many sites?) compared with the intensity (how many trees per site?). 

- 42 - 


Initially, sampling is done throughout the range of the species (Barner 1974; 
Turnbull 1975). If adequate information is not available on the distribution of the 
species, searches of literature and herbaria, and contact with foresters, amateur 
botanists, and others living in or close to the species range may help define its bound- 
aries. Aerial photographs, if available, can sometimes be used to save time in picking 
out possible collection sites in unknown country or in difficult terrain (Turnbull 1975). 

Ideally, taxonoroic and botanic exploration should precede collection, as efficient 
sampling schemes can only be devised based on knowledge of the distribution and ecology of 
the species. Sometimes, however, the activities of exploration and preliminary collection 
will have to be combined. This kind of a single, combined expedition cannot be expected 
to furnish all answers, and a series of reconnaissance and seed collection missions are 

The number of seed sources (provenances _!/) to be sampled will depend on the 
extent and the heterogeneity of the distribution range and on the genetic diversity of 
the species. As little will be known about variation patterns of the species in first- 
stage studies, sampling can be done on a fairly coarse grid, collecting at rather widely 
separated intervals following environmental gradients. 

Genetic diversity is generally largest in areas which are optimal for the develop- 
ment of the species in question. However, at the limits of the ecological range, outlying 
populations of a species may be exposed to extremes of temperature, rainfall or edaphic 
conditions. Such provenances may possess morphological and physiological characteristics 
which will be of great potential for particular environments. For this reason it is par- 
ticularly important that these marginal sites are included in the collections (Turnbull, 

When promising, broad general regions have been located through first-stage pro- 
venance trials, second-stage provenance collections should be concentrated on these, 
through sampling on a finer network (more collection points in a more limited area). 
Examples of sampling schemes used for some specific species are discussed by Turnbull 

The specific stands from which seed is collected will have to meet certain cri- 
teria. A stand has been defined as a sufficiently large population of trees possessing 
sufficient uniformity in composition, constitution and arrangement to be distinguished 
from adjacent crops (OECD 1971). In practice, the main consideration should be 'to select 
a population which is large enough to allow for sufficient cross-pollination between a 
large number of trees and which is isolated from related species to minimize risks of hy- 
bridization (Turnbull 1975; Melchior and Venegas Tovar 1978). The application of such 
criteria to some tropical species which are found as isolated trees in mixed forest may be 
difficult. In such instances, individual tree lots may be kept separate or combined with 
others covering a considerable area to represent f a stand' (Turnbull 1975). 

Melchior and Venegas Tovar (1978), referring to Eucalyptus globulus plantations 
in Colombia, consider 300 individuals a minimum number for a stand suitable for seed 
collection purposes. 

Stands should be of such an age that seed is produced in quantity, due to prac- 
tical as well as genetic considerations. 

In the case of introduced species, seed should whenever possible be collected from 
stands of known origin 2/. The general history of the stand is relevant in the case of 
both indigenous and introduced species. Any treatment which may have altered the distri- 
bution of the phenotypes should be noted and stands which have been selectively thinned 
to remove the best phenotypes should be avoided (Turnbull 1975). 

I/ Provenance: The place in which any stand of trees is growing. The stand may be 
indigenous or non -indigenous (OECD 1971). 

II Origin: For an indigenous stand the origin is the place in which the trees are 
growing; for a non-indigenous stand the origin is the place from which 
the seed or plants were originally introduced (OECD) 1971. 

- 43 - 


In selecting trees for sampling for provenance studies the aim is to take a sample 
which is as representative as possible of the population. The main considerations in 
sampling are the number of trees, the type of tree and the distance between trees to be 
sampled. In order to give maximum flexibility in selection it is desirable that seed 
collections take place when there is a good seed crop on the majority of trees (Turnbull) 
1975)* Good seed years will also furnish a seed sample which more fully represents the 
population genetically. 

Sampling at the individual level can be done either for variability or for superio- 
rity in specific characteristics. Provenance studies are designed to expose genetic 
differences between populations and to indicate the best localities for seed collections. 
Such research can be done as well, if not better, with seed from random trees as with seed 
from carefully selected phenotypically superior individuals. IUFRO standards suggest 
selecting 'average or not less than average 1 trees of dominant or co-dominant status in 
'normal 1 as compared with 'plus 1 stands. 

In sampling trees for provenance trials from natural stands, the question of 
spacing between sample trees is important because of the need to avoid trees which are 
closely related genetically(half-sibs) or trees which have an abnormally high incidence 
of self-pollination. Although recommendations vary, it is generally considered that seed 
trees should be between 100 and 300 m from each other to avoid narrowing down the varia- 
tion sampled due to relatedness or inbreeding (Turnbull 1975; FAO 1975; Melchior and Venegas 
Tovar 1978). Adjacent trees in plantations are usually not closely related because the 
seed has been bulked and so there need be no restrictions on sampling from adjacent trees 
(Turnbull 1975). 

The number of trees sampled per stand will vary according to the species and its 
breeding system. Ten to twenty-five trees can be considered a minimum in species" which 
are found in stands (Barnar 1974); possible solutions in dealing with tropical species 
which occur as individuals rather than as stands have been discussed above. 

The number of seed required per tree for provenance trials need not be large: 
10 000 seeds per tree would be sufficient for each experiment, provided seeds are picked 
out one per container or individually sown in seedbeds. 


The main consideration when sampling for conservation ex situ is the maintenance 
of maximum allelic diversity. Strategies are discussed in the lecture 'Principles and 
Sstrategies for the Improved Use of Forest Genetic Resources'. 


Collection for utilization is generally done from a limited part of the range, 
from populations which have been found through provenance trials to be well-adapted to 
the environmental conditions of the planting site. However, as many countries will start 
their planting programmes before sufficient information is available on provenance 
performance, climatic and edaphic matching is often done toselect the most likely seed 
sources. As most species and provenances have some degree of plasticity (i.e. capacity 
to adapt to environmental conditions which differ somewhat from the natural ones) 
(Willan 1979), this can be an acceptable temporary measure provided that full-scale 
evaluation trials are started in parallel. Systematic evaluation is always necessary, 
as other factors than climate and soil in the present environment will determine the 
genetic make-up of trees, making the performance of new introductions in a matching 
environment unpredictable. 

If collection is carried out in plantations or in even-aged, homogeneous stands, 
the genetic quality of the seed can generally be somewhat improved by sampling for superi- 
ority of stands as well as of individual trees (Turnbull 1975). However, it should be 
noted that in the case of uneven-aged natural stands and stands in which past history is 

- 44 - 

not known, phcnotypic performance is no guarantee of genetic superiority: the probability 
of superiority being due to purely environmental causes is, in fact, 50% (Melchior 1977; 
Barner 1974). 


The lowest acceptable standard of collection at the individual level 9 often used, by 
necessity f to meet the needs for substantial quantities of seed for large-scale afforesta- 
tion programmes, involves the collection of seed from all trees of a particular provenance 
or provenances t except from those of notably inferior phenotype (Turnbull 1977). Ideally, 
the trees sampled within a stand should be dominants, free of pests and diseases and in the 
case of plantations, of better than average form. 

In stands regenerated by natural seed fall, collecting from trees growing closely 
together may result in plantations with a narrow genetic base, thus responding in a uniform 
manner to environmental pressures such as diseases or unforeseen adverse conditions, and 
generally having less flexibility to adapt to the requirements of a new site. In addition, 
if seed stands are subsequently established in these stands or seed is collected from them, 
severe inbreeding effects may occur in the resultant second generation stands. In natural 
stands, the distance between trees collected should, as in collections for evaluation, 
ideally be 100-300 m and as a minimum requirement, trees should be selected at an interval 
that exceeds normal seed fall (Turnbull 1975). 

When collecting bulk quantities of seed there is no upper limit for the amount of 
seed collected per tree, provided that the number of trees collected is large, When 
collecting from standing trees, the lower limit of seed per tree would be set by economic 


Practical limitations and economic possibilities often modify the ideal sampling 
strategies outlined above. Factors such as accessibility of stands and yearly fluctua- 
tions in the seed crop at the provenance as well as at the individual level also influence 
decisions on sampling. If, however, the genetic and biological principles on which 
recommended sampling methodologies and intensity are based are known, these modifications 
can often be accommodated without too wide-ranging adverse consequences. Detailed record- 
ing of sampling procedures used and number of provenances, stands and individual trees 
collected from, as well as selection criteria in each instance, are of fundamental impor- 
tance and cannot be over stressed. 


Barner, H., 1974. Classification of Sources for Procurement of Forest Reproductive 
Material. In; Report on the FAO/DANIDA Training Course on Forest Tree Improvement, 
held in Limuru, Kenya, September-October 1973. FAO/DEN-TF-112. FAO, Rome 

FAO, 1975. Forest Genetic Resources Information No. 4. Forestry Occasional Paper 1975/1. 
FAO, Rome. 

Melchior, G.H., 1977. Program* Preliminar de un Ensayo de Procedencia de Cordia alliodora, 
Cupressus lusitanica y otras Especies Nativas y Exoticas. Proyecto Invest igaciones y 
Desarrollo Industrial Forestales COL/74/005. P1F No. 7. Bogoti, Colombia. 

Melchior, G.H. & Venegas Tovar, L., 1978. Propuesta para Asegurar el Suministro de 

Semi 11 as de Eucalyptus globulus en Calidad Comercial y Geneticamente Mejoradas. Projecto 
Investigaciones y Desarrollo Industrial Forestales COL/74/005. PIF No. 14. Bogoti, 

OECD, 1971. OECD Scheme for the Control of Reproductive Material Moving in International 
Trade. Paris, France. 

- 45 - 

Turnbull, J., 1975. Seed Collection - Sampling Considerations and Collection Techniques. 
In: Report of the FAO/DANIDA Training Course on Forest Seed Collection and Handling, 
held in Chiang Mai, Thailand, February-March 1975. FAO/TF/RAS-11 (DEN). FAO, Rome. 

Turnbull, J., 1977. Seed Collection and Certification. In: Selected Reference Papers, 
International Training Course in Forest Tree Breeding. Australian Development Assistance 
Agency. Canberra, Australia. 

Will an, R.L., 1979. Eucalypts for South-East Asia. Tropical Agriculture Research Series 
No. 12. Ministry of Agriculture, Forestry and Fisheries, Japan. 

- 46 - 

C. Palmberg i / and G.H. Melchior - / 



Introduct ion 47 

Biological pre-conditions for flowering and fruiting 47 

The juvenile-adult phase. 47 

Actual population size 48 

Planning seed collection 49 

Determining seed requirements 49 

Harvest forecast and evaluation % 49 

Sampling in seed collection 50 

Seed collection techniques 50 

On-ground collection 50 

Collection from standing trees 51 

Training and safety 53 

Packing and labelling collected seeds 53 

Drying, post-harvest ripening and seed cleaning 53 

Supervision 55 

Bibliography 55 

Annex 1. Equipment needed for seed collection, information on area and sample 

for herbarium 57 

Annex 2 and 3. Sample seed collection data sheet 58-59 

y Forest Resources Division, Forestry Department, FAO, 
Via delle Terme di Caracal la, 1, 00100 Rome, Italy 

2/ Bundesforschungsanstalt fur Holz und Forstwirtschaf t , 
Institut fur Forstgenetik und Forstpf lanzenziichtung 
Sieker Land Strasse 2, D-207 Grosshahnsdorf 2 
The Federal Republic of Germany 


A foremost constraint of large-scale planting of native or exotic species in Latin 
America, as in other parts of the world, is the lack of available information on basic 
theory and practical methods of harvesting, handling and treating forest seeds. Lack of 
contacts with the outside and especially lack of foreign exchange mean that the seed gap 
often cannot be closed by purchases abroad* 

Since the Latin American countries still have relatively extensive forested areas, 
domestic requirements for seeds of native species could be met by collections in local 
stands. There is an urgent need to establish seed areas or plantations to produce exotic 
seeds so that natural self-reliance in seeds can be achieved as soon as possible. 

Forest seeds may be collected to: 

- ensure the continuous short and long-term supply of reproductive material for 
plantation programmes. Seed collection to be determined by the demand for final 
products within the country; 

- guarantee the reproductive material needed for scientific trials; 

- supply the reproduction material for the establishment of ex situ gene banks, 
and for the establishment and expansion of a r bo re turns, botanical gardens and 
other collections of tree species; 

- supply reproductive material for beautif ication of landscapes, recreational 
sites, cities, highways and the like. 

Reproductive material collection methods vary in accordance with the above- 
mentioned objectives and will depend on the species in question, the sites to be re- 
afforested, the type of reproductive material to use, and so forth. 


Detailed discussions of the formation of gametes and zygotes in conifer and hardwood 
tree species can be found in the Mittak manual (1978) and in various biology texts. This 
lecture will therefore be confined to pointing out certain biological aspects of particular 
importance for the development of forest seeds. 


The zygote, seed and sporophyte are formed through the union of masculine and 
feminine gametes. A diploid individual is produced by mitotic cell division. This 
individual can be defined in terms of its specific morphological and physiological features 
such as: the position of the needles, shape of the leaves and duration of the juvenile 

The characteristic feature of the juvenile period is that the individual is incapable 
of sexual reproduction. The length oft the period is further influenced, in addition to 
hereditary factors, by environmental circumstances. The duration of the period can range 
from 2 to over 50 years, depending on the species. Cordia alliodora in lowland areas for 
instance, matures in 4-6 years, Podocarpus in mountain areas requires 30 years or more, and 
Quercus in the temperate zone over 50 years. 

When the tree reaches maturity its morphological characters change. The initiation 
of flowering and fruit setting is one criterion closely related to the degree of maturity 
of the tree. Unfortunately we know very little about phenological variation among and 
within tropical forest species and about the genetic and environmental influence* But 
it is known that individual trees within a specific stand or forest usually vary as to age 
at first flowering and the regularity of flowering and fruit setting, especially when young. 
One result of this variation, aggravated by the typically great distances between individual 
trees or groups of individuals of the same species in tropical forests, is the birth of 
subpopulations which all flower at the same time, if these subpopulations are small there 

- 48 - 

is a grave danger of self ing among them with all the adverse effects of inbreeding for the 
next generation: instability, low resistance to biotic and abiotic factors, and reduced 
vigour. This is why it is absolutely necessary to resist the temptation to collect seeds 
in very young stands, despite the fact that collection would often be easier and less 
expensive in such stands than in mature ones* Only when 60 to 100 percent of the indivi- 
duals in a given population have flowered (only adult populations will achieve such per- 
centages) should seed collection for commercial plantations be undertaken. 

To speed up the onset of the adult phase of the tree and thus speed up flowering 
and fruit setting, stands for seed collection are usually established in sites with fertile 
soils and optimum climates for the flowering of the specific species. Sometimes it is also 
possible to speed up the adult phase and increase flowering and fruit setting by applications 
of fertilizers or plant hormones. 

None of the above-mentioned methods should be used unless accompanied by careful 
observations or prior experimentation to throw light on the specific characteristics and 
requirements of the species. 


By population size is meant the proportion of genotypes actually participating in 
fertilization and seed production within any given year. 

Ideally, to enhance the physiological and genetic properties of seeds, all gametes 
in populations with a large number of unrelated individuals should combine at random. 
This is because crosses with foreign pollen are a positive factor in selection, as they 
maintain allele heterozygosity and suppress the appearance of sub-lethal alleles. 

The following factors affect actual population size: 
(i) absolute population size; 

(ii) mechanical barriers, which might consist of: natural windbreaks (reducing the 
movement of wind-borne pollen) or a very large mixture of non-hybridizing 

(iii) phenology and synchronization of flowering. Non-synchronized flowering can 

combine with scant flowering some years to drastically reduce actual population 
size and hinder continuous seed supply and progress in breeding programmes. 
Synchronization will affect the genetic make-up of the seeds collected, as 
genetic composition usually varies within the stand from one year to the next 
depending on which trees are flowering. It will often be necessary in 
breeding programmes to establish a series of seed orchards, each having a pheno- 
logically synchronized subpopulation; 

(iv) intra-genotype incompatibility. This phenomenon, which is usually a barrier to 
selfing, in particular, and fertilization among closely related trees can be 
based on the following factors: 

- the pollen does not germinate on the stigma; 

- the pollen tube does not penetrate the stigma; 

- the pollen tube interacts with the tissues of the style and does not reach 
the ovule; 

- the gametes do not function due to incompatibility; 

- gamete union takes place but the zygote dies as the endosperm is incompatible 
or else, in polyembryonic species, selection operates against the gamete. 

- 49 - 



Planning for reforestation must include seed procurement and ensure a sufficient, 
constant supply of seeds, 

The person in charge will calculate seed needs for forest nurseries or direct plant- 
ing in terms of the size of the projected plantation, the geographical area and the species 
one plans to use for reforestation. If the plan is to reforest 1 000 ha with Pinus oocarpa, 
at spacings of 2.5 x 2*0 m, for example, the number of plants required will be 2 000 
plants/ha. Therefore 2 000 000 plants will be needed to reforest 1 000 ha. Two 
good quality seeds are needed to produce one plant in a nursery bag in the forest nursery. 
The number of viable seeds needed for the planned reforestation will therefore be: 

2 x 2 000 000 = 4 000 000 

The number of viable seeds in a specific volume of cones varies greatly with the 
species, and even within the same species. Assuming an average 48 000 Pinus oocarpa seeds 
per hectolitre of cones, the number of seeds in the example quoted above is 

4 000 000 : 48 000 = 83.3 hi. 

It therefore follows that approximately 83 hi of mature P. oocarpa cones must be 
collected to reforest 1 000 ha (plants in nursery bags). 

Nonetheless, as it is unlikely the required number of seeds can be collected every 
year, the largest possible number must be collected in good years and kept in cold storage 
for the future. The provision for "x" number of years to be collected in a planned harvest 
will depend on such factors as flowering period and the storage life of the seed under 
existing storage conditions. It is a good idea to store enough seeds for at least two 
years plus the present year. In this instance, the amount of seed to be collected would be: 

3 x 83 = 249 hi 

Yearly records of production cone maturity dates for each species, and provenance 
are important, good planning and data on the periodicity of good harvest are dependent on 
such records. 


When the reproductive buds have formed, the collection planner or supervisor can 
use the appropriate sampling method to determine whether a potential harvest is develop- 
ing. We stress the word potential because the many impediments to seed development can 
destroy the reproductive structures or maturing seeds at any stage of the process. 

Early forecasts are based on sampling. Female buds are collected as samples from 
10-20 trees. Buds for identification are collected three branches from the top of the crown. 
A statistical evaluation of the sample material will give an indication of harvest potential. 
Theoretically, the greater the production of female buds, the greater the probability of a 
good seed harvest. Unfortunately, the quantitative relationship between female bud counts 
and potential harvest sice is not completely reliable. But comparisons between areas and 
periodicity 4re always of some practical value* The procedure described may be useful until 
such time as a more suitable technique is devised for forecasting harvests of a species 
important for reforestation. 

The forecast of subsequent forecasts may be based on a female flower count or, later, 
a count of new cones. 

I/ This part of the lecture is based on Mittak (1978). 

- 50 - 

The most common error in classifying cone harvest! is to count old cones which have 
scattered their seeds in earlier seasons, and to use trees growing along roads as a basis 
for estimates. Such trees receive more sunshine and frequently have more cones and buds 
than those within closed stands. 

Based on the qualifications of the various possible seed collection areas, the planner 
will assign collection priorities subject to the next seed yield evaluation (content) of 
the harvest. 

The main objective of the harvest evaluation is to indicate the number of healthy 
seeds in cones. Cones are picked at random from different trees throughout the stand. Care 
must be taken to ensure that the samples are representative of all cones, i.e., they must 
be collected evenly all over the crown of each seed tree. 

Some ten representative trees in the stand are selected for evaluation and ten cones 
collected from each. The ensuing one hundred cones are split lengthwise down the middle for 
examination. All the good seeds on one of the cut surfaces are counted and then minutely 
examined. They are then treated separately in a 65C oven until the pine cone sections 
open releasing the seeds. The total number of good seeds from the cone is divided by the 
number of seeds counted on a single face of the split cone. An average figure can be 
worked out for the one hundred cones once these figures are available. 

Seed pests and diseases not only affect cone or fruit ability to produce mature seeds, 
they also affect subsequent seedling development. The usual estimate is that if more than 
50 percent of the seed is damaged, seeds should not be collected from that stand. 

Together with the quality and number of seeds it is important to get some indication 
of how ripe they are so a date can be set for harvesting. 

The seed incision test, used in conjunction with other indicators of maturity, is 
usually very helpful in determining maturity. The test consists of cutting each seed 
lenthwise exactly down the middle with a razor blade. The contents of 20-30 seeds are 
then examined under a 10 power lens. The normal process is that as the seeds mature, the 
embryos elongate and turn yellowish whilst the endosperm turns from milky and sticky to 
firm consistency (similar to coconut meat). The tegument and wing also darken. Usually, the 
embryo has to have elongated at least 75 percent of its potential length to ensure seed 
viability. The possible or potential length of the embryo is the length of the cavity 
within the endosperm. 

The incision test should not be performed until 3-4 weeks before seed maturity as 
before then the unfertilized ovules of some species can simulate normal development. In 
other words, they contain endosperm and look like good seed. However, since they have not 
been fertilized, there is no embryo and they cannot form viable seeds. Some three weeks 
before maturing, most of the unfertilized ovules stop developing and their endosperm tissue 
will dry up and form empty pods. 

See previous lecture. 


Seed collection from natural seed fall, being a low-cost, easy method, is very 
widespread. It includes collecting seeds of fruits which have fallen from standing trees 
and collecting from felled or fallen trees* 

This method is frequently used for fairly large seeds or fruits, including, in 
temperate regions Quercus spp., Fagus spp., Castanea spp.; in tropical areas Tectona spp., 
Sb&Cfii spp* Triolociiiton spp. and Gmelina arborea (Turnbull 1975; USDA 1974). Where 
possible it is recommended that the ground under the trees be cleared prior to collection. 

- 51 - 

Although seed collection from natural seed fall is a relatively easy, economical 
method, it does have certain serious drawbacks. The seed viability of many species such 
as Shorea spp., is very quickly lost once the seeds have been shed. Seeds on the ground 
are also very prone to damage by insects, fungi and animals. It is therefore imperative 
that shed seed be picked as soon as possible, remembering that the first seeds or fruits 
to fall are usually of rather poor quality (Turnbull 1975). Another drawback with this 
method is that it is normally not possible to say from which tree the seed was shed and 
there is, therefore, no indication of phenotype quality. 

A closely related method which does eliminate many of the above drawbacks is to use 

tree shakers. These consist of a hydraulic arm mounted on a tractor which is used during 

the collection season to shake the trunks to make the ripe fruit fall (Turnbull 1975; 
Ottone 1978). 

Animals such as squirrels sometimes gather cones or seeds. Foresters sometimes raid 
these stores for quick 'seed collection* It is a very common method e.g. for Pseudotsuga 
menziesii in the United States (Turnbull 1975). 

Collecting seeds from felled or fallen trees in thinnings or clear-cut areas or 
from trees grown specifically for seed collection is another common harvesting method. 
Though in theory it is possible to limit the collection to desirable stands using this 
method, one can never again collect seed trom tne same tree or stand. Cutting down trees 
just to collect seeds in shortsighted and wasteful. 


Seed collecting from standing trees is the most common large-scale method for 
collecting forest seeds. It is a sure method, but careful attention must be paid to follow 
the proper safety measures, use the right equipment, and' keep the equipment in good 

They are two distinct harvest methods (i) collecting from standing trees without 
climbing, (ii) collecting from standing trees by climbing them. 

(i) Collection from standing trees without climbing 

There are various ways of harvesting seeds without climbing. Some use equip- 
ment to get the fruits and seeds down, such as long, light poles for striking 
and shaking branches, knives with telescopic or long handles for cutting fruit, 
collection sticks, pole proners and the like. 

In Australia a rifle is used to sever the fruit-bearing branches of trees. 
Another possibility is a flexible saw operated by two collectors from the base 
of the tree. This last method can be used to cut branches up to 20 cm in 
diameter (Turnbull 1975). 

(ii) Climbing standing trees for seed collection 

Tree climbing equipment varies with the species and environmental surroundings. 

Burley and Wood (1978) lists of equipment needed for seed collection are found 
in Annex 1. 

Climbing irons 

The usual way to climb trees is to use climbing irons. The forged iron is fastened 
to the foot gear with two sturdy leather straps* The foot gear must be solid and firmly 
attached to the climber's foot and leg. The iron ends in a fixed spur of varying length 
depending on the tree climbing method used. One of the best has a short hook which does not 
extend past the sole of the boot. This enables the climber to walk on the ground without 
difficulty. Irons can damage trees with thin smooth bark and should therefore not be used 
for climbing young or thin barked trees (Mittak 1978; Turnbull 1975). To facilitate the 
climber's task, ropes can also be snaked around the trunk of the tree (Mittak 1978). 

- 52 _ 

The IUFRO team* collecting North American conifer seeds used as standard equipment 
climbing irons, safety belt, a hook to bend back branches, safety hat and overalls 
(Turnbull 1975). 

Tree bicycle"Baumvelo" 

The use of a Baumvelo or tree bicycle is recommended for collecting seeds, cones or 
fruits from trees with very thin, smooth bark. Tree bicycles are very easy to use on trees 
with smooth boles. The tree bicycle is particularly used to collect buds and pollen from 
plus trees to avoid damaging them (Nittak 1978). 

The bikes consist of two pedals supported by a vertical arm. The pedal is fastened 
by special straps to the operator's boots. The arm is connected to a steel band forming a 
circle of adjustable diameter around the stem of the tree to be climbed. The pedals must be 
about 5-8 cm greater in diameter than the tree for ease of movement. In climbing, the 
operator puts all his weight on one boot-pedal, then raises the other foot with the other 
boot-pedal as high as possible. He then moves the other foot up to meet it. Between the 
two operations, he roust adjust his safety belt as high as possible. When he reaches the 

crown he can take off the pedals and fasten them to the trunk for freedom of movement. He 
must again strap on the pedals to climb down the tree. The tree bicycle is easy to 
transport. Climbing with the tree bicycle is easy and rapid after some practice (Mittak 
1978; Turnbull 1975; Ottone 1978). 

Portable Ladders 

Ladders for collecting forest seeds vary greatly in construction and the materials 
used. Ladders are a quick, safe method to climb trees up to 15 or 20 ra or even higher. 
Mittak (1978) lists the following kinds of ladders for collecting forest seeds: 

Rope Ladders, such as sailors use. A rope is shot over a sturdy branch to haul up 
the rope ladder which is then made fast to the branch. 

One-legged ladders, with alternate short rungs on both sides of the supporting pole. 
These ladders are appropriate for uneven terrain and branchy trees. One-legged ladders are 
attached by chains to the trunk of the tree. There is usually a bolster between the tree 
and the pole ladder at chain height to facilitate access. They are easier to steady than 
regular ladders. 

Scaling ladders in several sections, usually of aluminium, are needed for climbing 
tall trees. Sections vary in length (2-4 m), but for easy carrying no section should 
weigh more than 3-4 kg. The sections are designed to fit into one another and are attached 
by a chain to the trunk of the tree, which the operator grasps as he climbs. 

In addition to the above-mentioned ladder types, ladders mounted on tractors are 
used, as well as hydraulic platforms of various types. Often these platforms are only 
used in seed orchards since per unit collecting costs are very high (Turnbull 1975). 

Nets and other various methods 

A special device for climbing trees, especially small trees such as cypress, is the 
net built by the United Kingdom Forestry Commission. A triangular net is suspended by 
special ropes and snatch blocks from the crown of the tree. The operator can climb up the 
net to collect the small fruits which cannot otherwise be harvested without partially 
cutting branches, doing great damage to future harvests. This device produces good results 
in species like cypress, i.e. species with many small cones or fruits on the outer part of 
the crown (Mittak 1978). 

Mention has also been made of the use of pulleys attached to a sturdy fork. Even 
balloons have been used in different parts of the world, but they are hard to use and the 
end result is generally disappointing (Ottone 1978; Turnbull 1975). 

- 53 - 


IB is very important to train climbers in the various phases of seed collection and 
the correct use of tools and equipment. 

Foremen should know which trees will produce good seed and which will not, and mark 
them prior to harvest. 

Personnel safety is, after genetic and biological know-now, the most important aspect 
of seed collection. 

Personnel should receive practical and theoretical training in accident prevention 
and first aid. 

Each team should be equipped with a first aid kit and stretcher. Tools, safety 
belts, safe hats, protective goggles, ropes, hooks, spurs, ladders, the first aid kit and 
all utensils for seed collection should be minutely examined before the climb or before 
beginning work. Any tool or other equipment which gets damaged on the job should immediately 
be reported and a request made for repair or replacement. 

Equipment should be clean. A product should always be brought along to dissolve 
resin and other sticky substances like pitch once the Job is finished. Equipment 
maintenance promotes safety and makes the equipment last longer. 


After picking, dry cones should be packed in coarse sacks or in solid-bottomed fruit 
crates. It is not advisable to use plastic bags for this type of cones. Air circulation is 
inadequate and moisture condensation can affect the fruit and seed. The temperature also 
rises within plastic bags, which can cause various kinds of damage such as lowered germina- 
tion rates, mildewed seeds and fewer seeds released. 

The coarse sacks are not completely filled. They are tied with string but not sewed, 
so the work will go faster. The reason the sacks are not completely filled is to avoid 
tightly packing the cones while they are still in the field and during the pre-drying 
period to prevent mildew from moisture and lack of air circulation. The sacks should be 
protected from rain and rodents and covered with canvas tarps. The sacks should be moved 
and turned two or three times a day and should not touch* The ground or floor should be 
dry. If crates are used, they can be piled up to a height of 2-3 m. The corners at the top 
of the crate and slatted sides promote air circulation. This stage of the process 
constitutes a sort of pre-drying phase affecting subsequent work and final seed quality* 
The cones can also be more easily moved to the processing site. 

Sacks and crates are also a way to keep the cones during the collection/drying 
period* The harvest period is short and so a large number of cones is being handled. Thus 
packed, they will keep perfectly (Ottone 1978). 

Some species of fruits lose viability if not kept moist. They must be collected 
before dry and transported to the temporary storehouse as soon as possible. Plastic bags 
can be used for some species for very short times. Sturdy paper bags with or without an 
inner plastic lining are also used (USDA 1974). 

To guard against lost identity of seed lots it is extremely important to identify 
and label them. Each sack or package should be labelled inside and out. The labels must 
be water- and moisture-resistant and will list data on the species, provenance, collection 
date, name and last name of the collector and so forth. A seed lot number identifies them 
on a more detailed data sheet (see Annexes 2 and 3 for sample sheets). 


The cones or fruits should be transported as quickly as possible from the temporary 
storage area to the final handling and storage centre* 

17 ThU part of the lecture is based on Mittak (1978) 
|/ This part of the lecture is based on Ottone (1978) 

- 54 - 

The first step after transport is to dry the cones and extract the seed. There are 
two mains drying methods: sun and air drying, and artificial drying. 

The natural drying process involves less risk of lowering seed quality but takes 
longer. Nor can the process be used in very wet or rainy climates. For easier handling 
of cones and seeds and quicker cone drying, the cones and seeds can be piled onto canvases. 
The canvas is folded at night to protect the cones and seeds from dew or rain. Rough sheds 
in which the air circulates freely are another answer. Another possibility is trays which 
are shaken to make the cones release the seeds. Plastic roofs can accelerate the sun 
drying process. Tin roofs are not used as they cause moisture condensation. 

Increased air humidity in the drying area can produce a condition known as case 
hardening ( the same can occur in oven drying). What happens is that the cone reabsorbs 
water and the opening process is inhibited to the point that the cone closes or remains 
partially closed, preventing seed release. 

Humidity and temperature are easier to control with artificial drying. Artificial 
drying is quicker (from 6-15 hours) but it does require equipment and installations which 
can be very costly. It should be considered that the equipment is only used a few days of 
the year. Artificial drying is only recommended for cold, wet areas. Drying should be 
done in the shortest possible time to avoid damage to the seeds. The air must be hot and 
dry for even, rapid drying. Case hardening can occur if the cone is reheated. Cones must 
be pre-dried. Good oven temperature control is essential. The cones should be dried at 
the minimum temperature for as short a time as possible. Ideal temperature range is 10C 
at the beginning of the process and no higher than 52 W C at the end. 

High temperatures will cause physical and physiological damage to the seed as will 
humidity (which should be eliminated as quickly as possible). The p re-dry ing operation aids 
and shortens oven drying. Pre-drying is therefore always recommended - for how long 
depends on the tree species. If rotary ovens are not used, once the fruits have dried they 
can be put into a 60 x 150 cm wire drum ( approximate measurements), which is rotated to 
detach the seeds not released during the drying process. The same should be done with sun- 
dried cones. Small quantities of seeds can be threshed by hand into a crate, or onto a 
tarpaulin or brick or wooden floor. 

Eucalypt capsules release their seeds in a few hours or at most one day after 
picking. The best way to collect eucalypt seeds is therefore to spread them on canvas 
tarpaulins sheltered from rain. The seeds can easily be separated from the fruit, branches 
and leaves by straining them through a fine mesh screen. They are then packed in closely* 
woven cloth sacks. 

Fleshy fruits usually have to be soaked to remove the external fleshy covering. The 
seed is then removed by hand or with special equipment. The treatment includes alternate 
phases of washing, drying and cleaning and must be begun immediately so the fruits will 
not have time to ferment (USDA 1974). 

One limiting factor in collecting the fruits of various forest species is the time 
period in which harvesting is possible. The cones or fruits of some species can be picked 
before ripe to lengthen the collection period. They must be stored in such a way that they 
can complete their anatomical and biochemical development. The usual post-harvest ripening 
temperature, depending on the species, should be 5-1 8C. Constant vigilance is necessary 
throughout the process to ensure that the above conditions are met and no problems arise 
which might adversely affect the fruit and seeds, such as mildew. 

It is also important to know how early the fruits are to be picked. Certain pine 
species can be picked when they have reached 1.1-1.2 of their specific weight, or approx- 
imately 15-20 days before the normal collection date. 

The fruits or cones are stored as above for one month during which time they complete 
the development process. They are then dried in the usual way. The wings of winged seeds 
must be removed after drying. 

To remove the wings the seeds are placed in a wire drum containing two or three sets 
of rotary brushes, regulated in accordance with the size of the seed to be processed. These 
brushes are attached to a central axis driven by an electric motor. As the drum turns, 

- 55 - 

the seeds and wings are pressed against the wire and the wings are rubbed off. 

The brushes must also be set to leave enough space between the brush and the mesh 
to avoid excessive rubbing which would raise the temperature, harming the seed. 

If winged seeds are to be sown immediately after, they can be slightly moistened. 
The wings will swell with the absorbed moisture and can easily be removed. If seed treated 
in this way is to be stored, it must be dried before storage. 

Ordinary wire sieves with wooden frames can also be used to remove the wings from 
the seeds. These are vegetable seed cleaners with mesh sizes calibrated to seed size. 
Seeds can also be winnowed directly out-of-doors or in variable draft winnowing fans. 

Whatever method is used for winnowing, the operation also eliminates sterile seeds. 
Seeds should not be over-winnowed to avoid eliminating good seeds which happen to be lighter 
because they are smaller. It would be better to keep some sterile seeds than to lose good 

At all stages of handling cones, fruits and seeds, it must be remembered that they 
are living organisms. Rough handling will lower germination rates and adversely affect 
seedling development* 


Before leaving the field, the supervisor and foremen must check all the equipment. 
In addition to checking the equipment, the supervisor is also in charge of transport, fuel 
and daily subsistance allowance for his team. A successful harvest, staff safety and the 
cost of seed collection are largely dependent on how well the operation is organized. 

Every day the supervisor marks the trees in the area from which seed is to be 
collected. The criteria is seed maturity. 

It is recommended that harvesting be done by work quotas, first determining costs in 
accordance with the amount to be harvested and reception facilities. Inspectors should en- 
sure strict compliance with quality standards. 

The supervisor must also keep a written record of the output of individual collec- 
tors, listing the number of cones, seeds or fruits collected, the operation, the species, 
provenance and site number and any other necessary observations. He must also fill out the 
collection sheets (see Appendixes 2 and 3). 

It is recommended that a random check of one sack in five from each collector be made 
to ensure that the fruit or seed collected has been properly identified, that the sack is 
not over-full and that good-quality seed has been picked. 

Supervision is essential at all subsequent phases of handling to avoid damaging the 
seed and ensuring that each seedlot is properly identified. Any treatments must be listed 
on the corresponding sheets for each lot. 


Burley J. & Wood, P.J., (1979). (Editors) Handbook on Species and Provenance Research 
with Special Reference to Tropical Areas, Commonwealth Forestry Institute. Tropical 
Forestry Papers No. 10 & 10A. Oxford, U.K. 

Mittak, W.L., (1978). Manual 2. La recolecci6n de semi lias forestales. INAFOR BANSEFOR 
FAO/TCP: Proyecto GUA 6/01-T Institute Nacional Forestal, Guatemala Ciudad. 

Ottone, J.R., (1978) Recoleccion y tratamiento de frutos para obtener semi 1 Us forestales. 
Institute Forestal Nacional. Folleto Tlcnico Forestal No. 45. Buenos Aires. 

T7 This part of the lecture is based on Mittak (1978) 

- 56 - 

Turnbull, J.W., (1975). Seed Collection: Sampling considerations and collection tech- 
niques. In: Report on the FAO/DANIDA Training Course on Forest Seed Collection and 
Handling, held in Chiang Mai, Thailand, February - March 1975. FOR . TF-RAS / 1 1 ( DEN ) . 

US DA, (1974). Seeds of Voody Plants in the United States, Chapter V. U.S. Department of 
Agriculture, Forest Service. Agriculture Handbook 450. Washington D.C. 


- 57 - 

Annex 1 


A - Seed collection 

Containers for seeds (in field). Sacks and bags (may "be re-usable). 
Containers for seeds (for shipping). Cotton and canvas bags (sent with seeds). 
Tags for trees, e.g., plastic rings. 
Climbing equipment. Climbing irons, tree bicycles or ladders. Safety belt, 

safety harness, safety helmets, tool -carrying harnesses. 
Seed cutters, e.g. hooks and rakes for cones, pruning poles, hand shears. 
(Heavy) plastic tarps to protect fruits, extract seeds and so forth. 
Binoculars to examine tree crown, check fruit development, etc. 
Portable walkie-t-t alkie (special permit may be necessary). 
Insecticide and fungicide dusts for seed protection (use with caution). 
Axes, saws, machetes, knives. 
Harnesses, ropes, labelling equipment and labels. 

B - Site description 

Notebook, site description sheets. 

Maps (including copies with major outlines to be filled out in detail). 



Meteorological equipment (hygrometer, maximum/minimum thermometer). 

Soil survey equipment (core drills, soil colour charts, pH test). 

Tree measuring equipment (altimeter, diaxnater belts, bark calibrators, etc.). 

Camera and equipment (wide angle lens). 

Tape recorder (batteries). 


C - Sample collection 

Plant pressing equipment. 

Packing material (local newspapers can be used). 

Plastic bags. 

Bottles for liquid samples. 

Preserving fluids. 

Growth corer (for wood samples). 

Carpenter's brace and drill (for resin samples). 

Insulated container (e.g., ice box). 

Magnifying glass. 

Insecticide spray (for herbarium mater ial). 

ALSO: Any needed medicines, camping equipment, vehicles and equipment as needed. 

I/ From Burley and Wood (1979) 


Annex 2 



Scientific name: Provisional number: 

Latitude : 

Longitude : 

E Altitude : 


Map reference: 


Province : 

Region and/or administrative unit: 



Soil type: 


Drainage : 

Monthly rainfall distribution: 


Plant association: 

Density: open 

Height : 


Stand condition: 


Number of trees: 


C Amount of seed/cones: 

Annual rainfall: 

dense Regeneration method: 

Date of collection: 
Spacing of trees: 
Condition of seed/cones: 

Potential for commercial scale collection: 


S Extraction method: 




Yield per unit of volume: 

Description written 

Treatment : 


- 59 - 

Annex 3 



Species: Pinus oocarpa Shiede Seed No.: K31 Storehouse No. 1/71 
Country: Nicaragua Province: Nueva Segovia 

District: Dipilto Area: El Junquillo 

Latitude: 13%2'N Longitude: 8635 ! W Altitude: 1000m 

Situation; In western Nicaragua on the southern slopes of the Dipilto Cordillera. The 
Cordillera forms the northern boundary of Nicaragua in this area. The stands, which are 
some 5 km north of Macuelizo and 8 tan west of Dipilto, are part of an area of some 150 000 
ha of open pine forest which extends more than 70 km east and west along the cordillera 
and northwards to the pine forest of Honduras, Rainfall gradually increases eastwards and 
in this direction the lower slopes (below 800 m) gradually fill with populations of P. 
caribaea, whereas on slopes above 1 500 m P. pseudostrobus and deciduous forests tend to 
dominate. On the astern side near Macuelizo, the low slopes and valleys support only 
dry thorn bush, and ?. oocarpa is the only pine species. This is the major pine species 
throughout the cordillera. 

Soil; Sandy, quick-draining soils or pebbly soils with abundant quartz, derived in situ 
by wearing of granite rocks, with frequent outcroppings on the steepest slopes and peaks. 
Active erosion. Soils are usually shallow except in depressions or valleys, and contain 
hardly any humus. pH = 5-7 

Climate: Mean annual rainfall in Macuelizo (5 km to the south) is 90^ mm with the following 
distribution (in mm): 


21 16 113 180 110 85 1U9 163 1*9 13 

No temperature data is available for Macuelizo but at the Ocotal weather station (20 km 
east and kQQ m lower than El Junquillo), mean monthly maximums in the dry season range 
from 28-32C. 

Description of stand; Very open pine forest on very steep slopes (25-35) with very scant 
herb cover, including Andropogon spp, and Pennisetum spp. Some Quercus spp. are found in 
very wet valley areas. The largest pines are over 30 m in height with a diaaneter at 
breast height of 80 cm. Pine regeneration is scattered but good in various places. Tree 
rings are not very well-defined and are difficult to interpret, but it seems that growth 
is slow, from 3-5 rings per cm. 

Cone-bearing trees ; DBH UO-60 cm Height : 25-30 m 

Branch angle: 70-80 Bole: Straight, cylindrical, 


Collection methods; Selected trees on cutting sites (36 trees) 
Collection date: January, 1971 

- 60 - 


B. Ditlevsen 
National Forestry Service, Denmark 



Introduction 61 

Storage 61 

Principal factors affecting viability 61 

Moisture content 61 

Temperature 62 

Oxygen and other factors 62 

Special factors affecting viability 63 

Ripeness of the seed 63 

Fungus, bacteria and insects 63 

Mechanical damage 63 

Cytological and genetic changes 64 

Storage methods 64 

Seeds of natural longevity 64 

Seeds that can be stored 5*10 years with low moisture content and at low 

temperature 64 

Seeds that can be stored 3-5 years with moderate moisture content and at low 

temperature 65 

Seeds that can be stored 1-3 years with high moisture content and at low 

temperature . . . 66 

Very short-lived seeds 66 

Tests 67 

Sampling 67 

Purity analysis 67 

Viability tests 67 

Germination analysis ... ... 68 

Indirect viability tests 69 

Results of viability tests and their use 70 

Analysis of moisture content 70 

Determination of seed weight 70 

Health tests 70 

- 61 - 


Seed testing equipment 71 

Certification of forest seeds 71 

Certification programmes 71 

Implementation of certification programmes 71 

Bibliography 73 


Effective planning and production of afforestation programmes depends on the avail- 
ability at all times of sufficient quantities of seeds with the right physiological and 
genetic characteristics. In the first place, the seed must be stored until required, 
without losing its germi native capacity. Secondly, there must be continous checking, 
by testing the physiological characteristics of the seed. Finally, it is important for 
the future production of plantations that the material have the right genetic quality. 


Storage may be defined as the conservation of live seed from harvesting to sowing 
(Holmes and Buszewicz, 1958). 

The reasons for storing forest seeds may be summarized as follows (FAO, 1955): 

1) To conserve seeds in the best conditions to protect their germinative capacity 
from harvesting to sowing; 

2) To protect the seed against destruction by rodents, birds and insects; 

3) To preserve quantities of seeds harvested in good years, in order to have 
reserves available for years when little or no seed has been produced. 

Many forest species only produce seeds in sufficient quantities at intervals of 
several years; long- term storage is extremely important for a regular supply of forest 

It is usually assumed that respiration and metabolic activities must be considerably 
reduced if the seed is to survive a long period of storage. In practice, this means that 
only seed which can tolerate a reduction in its moisture content, and storage in these 
conditions, can survive a long storage period. It is also important that the seed be fully 
ripe before drying and that it not suffer damage during harvesting and handling. 


Viability may be generally defined as the capacity to survive or continue develop- 
ing. A live seed is one that is capable of germinating in favourable conditions. The 
germination of the seed is not necessarily either easy or rapid, and dormant live seeds 
may require lengthy special treatment to germinate (Owen, 1956). The roost important 
factors influencing seed viability during storage are the moisture content of the seed 
and the temperature. In some cases it has been shown that various gases in the surrounding 
area influence the seed somewhat (Owen, 1956). 


According to the rules of the International Seed Testing Association (ISTA), the 
moisture content of the seed is expressed as a percentage of the weight of the seed in 

- 62 - 

its original moist state (ISTA, 1976): 

Percentage of moisture = Weight of moisture x 100 

Weight of moisture + 
Weight of dry matter 

A reduction in moisture content considerably slows the metabolic processes, which, 
in turn, slows the respiratory process and the consumption of stored nutritional matter. 
However, it must be pointed out that some forest species with large seeds - for example, 
many leafy species - do not normally survive drying. It is also important that they be 
dried with great care. 

The moisture content of the seed increases before ripening. After ripening it is 
much lower. The natural moisture content in seeds after ripening varies between 15 and 50 
percent, depending on the species and on conditions. 

A change in moisture content during storage as a consequence of storage outdoors, 
or due to the opening and closing of the storage containers, destroys the capacity of the 
seed to germinate (Barton, 1961). 

The moisture content obtained in storage depends to a large extent on the relative 
humidity of the air inside the storage area. After a few days or, in some cases, a few 
hours a characteristic balance is reached between the moisture content of the seed and the 
relative humidity of the air (Holmes and Buszewicz, 1958). Changes in the moisture content 
depend on the relative humidity of the air, the moisture content of the seed and type of 
testa and size of the seed lot. 

It should be borne in mind that the results of studies of moisture content of a 
seed lot are mean values, and that there could be major variations in moisture content 
from seed to seed. Schonborn (1964) demonstrated, for example, that in one lot of Pinus 
sylvestris seeds with an average moisture content of 6.7 percent, individual percentages 
varied from 4.1 to 9.3. 


Seeds usually keep better at relatively low temperatures. Variations in temperature 
have an adverse effect on seed quality. 

The interaction between temperature and seed moisture content is very important in 
storage, and it is often difficult to separate the two factors* As a general rule, it may 
be said that when the temperature is low, the critical moisture content is higher than when 
the temperature is high. In other words, low temperature can to a certain extent compen- 
sate for high moisture content, and vice versa (Holmes and Buszewicz, 1985). 

Temperature influences absorption of moisture by the seed during storage. A general 
rule is that moisture content of the seed at a given relative humidity drops when the 
temperature goes down. 

Schonborn (1964) made studies to demonstrate the lowest temperatures that could 
be tolerated by seeds of different forest species at different moisture contents. 

Results show that seed tolerance of low temperatures is proportionate to low 
moisture content. Also, it appears from the studies that species such as Pices abies, 
Abies alba, Pinus Sylvestris, Fagus sylvatica, Pseaudotsuga, Betula and Quercus tolerate 
temperatures as low as -20C at a moisture content between 8 and 10 percent. 

For seeds requiring a high moisture content in order to conserve their gerrai native 
capacity, storage at temperatures below zero leads to frost damage, with the consequent 
loss of germi native capacity (Holmes and Buszewicz, 1958). 


The respiratory process depends naturally on the moisture content in the seed. The 
purpose of lowering temperature and moisture content during storage is precisely that of 

- 63 - 

lowering respiratory intensity, since this is the only way of ensuring that the seed main- 
tains its germinative capacity for a long time* 

It is of course desirable to store seeds with low moisture content in airtight 
containers that maintain a constant humidity, slow the respiratory process and protect the 
seed against pests. For seeds requiring a high moisture content during storage, such as, 
for example, Quercus spp., storage in airtight containers can be harmful, apparently due 
to the fact that this group needs some change of air during storage (Wang, 1977). However, 
it is clear that air composition within the container undergoes modification with time, 
due to respiratory processes, something which could influence the life of the seed, 

A series of studies has been made of the influence of different gases on the life of 
the seed, and the results have been published by Owen (1956) and Barton (1961). 


How ripe seed is at harvest is of course an important factor, and is responsible 
for some variations in seed viability. The timing of harvesting is therefore highly impor- 
tant, and it is also important to be able to determine the earliest stages at which a large 
quantity of live seeds can be collected. 

The degree of ripeness can be assessed by a series of different methods, described 
in detail by Earner (1975). 

However, it is often difficult to decide the most appropriate moment for harvesting 
seeds, especially in tropical countries where the period from maturation of the seed to 
shattering is often very short. In other cases there can be various degrees of ripeness 
on the same piece of land, or even the same tree, which can mean that unripe seeds have 
to be harvested, in the hope that post-harvest ripening can be induced before storage. 


Seeds stored under relatively humid conditions are easily attacked by the so-called 
storage fungi, of which the major groups are Aspergillus, Botrytis, Rhizopus and Penicillium. 
Holmes and Buszewicz (1958) provide more references. 

The most widely used method of avoiding attacks by fungi and bacteria during storage 
is the use of relatively low temperatures or relatively low moisture contents, under which 
fungi and bacteria cannot survive. The use of both a low temperature (maximum +5 C) and 
a low moisture content (maximum 10%) provides the best natural protection. 

According to Christensen (1972), the use of fungicides during dry storage often poses 
problems, since many fungicides have to be dissolved in water. 

It is unusual for seeds which were not already infested before storage to be destroyed 
by insects, especially Megastigmus species. 

In most cases, such losses can be avoided by controlling the temperature, bearing in 
mind that almost all insects in stored seed die at temperatures above 40-42C (Holmes 
and Buszewicz, 1958). Chemical products can also be useful in controlling insects; however, 
it should be borne in mind that these products can considerably reduce germinative power, 
particularly in the case of seeds with a relatively high moisture content (Ezumah, 1976). 


Mechanical damage can be defined as harmful changes due to damage during harvesting 
or handling of seed. Since the effects of the damage increase with time, mechanical damage 
also covers heavy damage resulting from mechanical damage (Moore, 1972). 

The reaction of seeds to damage varies very much, and some species have a greater 
natural capacity for recovery. Major damage will immediately lower seed viability, while 

- 64 - 

the effects of less important damage will usually only become apparent after a certain time. 
Consequently, it is important that seeds which are going to be stored for a long time 
should be treated with care during both harvesting and subsequent handling. 

The moisture content of the seed is important because of its effect on resistance 
to mechanical damage. Moist seed tends to swell, whereas dry seed tends to break. 

The best method of detecting the presence and kind of damage is tetrazolium testing 
(Moore, 1969). Radiographic methods such as those described by Kamra (1967) also produce 
good results. 


Lowered viability means that yield is reduced in two ways. First, a drop in the 
germination rate lowers the number of plants per area unit, and second, the surviving 
plants may be of inferior quality (Roberts, 1972) . 

Many experts fear that a considerable reduction of germinative power involves 
changes in the genetic composition of a seed lot since the survival capacity varies from 
one genotype of another (Harrington, 1970, Frankel, 1970). They also fear that poor 
storage conditions may increase the number of chromosome changes (Harrington, 1970). 

Wang (1977), however, reached the conclusion that these two kinds of genetic modi- 
fication can largely be controlled and kept to a minimum by efficient harvesting, handling 
and storage of seed. 

In practice, this means that seeds, stored under good conditions and which have 
maintained their germinative capacity can safely be used. But it is not very safe to 
use seeds with severely curtailed germinative capacity, when seeds have been stored for 
long periods, they should be used with circumspection, particularly in breeding and gene 


As many be seen from the above, there are great differences between different species, 
or rather, between different genera, as regards capacity to survive storage. Seeds can be 
divided into five major groups in the light of this capacity. Guidelines as regards 
storage methods for the different groups, and examples of genera and typical species within 
each group, are given below. 

1) Seeds of natural longevity. This group included Acacia, Robinia, Albizia, 
Sophora, Cere is, Cytisus and Gleditsia, all legumes, whose seeds have a hard 
testa and normally a very l ow moisture content. 

The seed should be stored in a dry place; the temperature, however, is not very 
important, and it is not necessary to use airtight containers. In some cases it 
is necessary to use disinfectants and other control measures. 

2) Seeds that can be stored 5-10 years with low moisture content and at low 
temperatures. Typical examples of this group are Picea and Pinus . 

Preparation for storage: The seed should be handled properly from the beginning, 
and should not be harvested until it is fully ripe. The cones should be kept 
cool, and damage to the seed and abrupt changes in moisture content and tempera- 
ture should be avoided. There should be adequate ventilation during extraction, 
and temperatures should not exceed 20 C at the start and 40 C at the end of 
extraction. After extraction the moisture content is very often 5-8 percent, and 
the seed should be dewed (trimmed) and cleaned as soon as possible before being 
stored. At the above-mentioned temperatures it is not normally necessary to 
disinfect the seed. 

Drying process* The moisture content of the seed should be ascertained as soon 
as possible, by one the methods described in the paragraph on seed analysis. 

- 65 - 

If the seed does not have the right moisture content, it must be dried. Drying 
can be in the sun, in hot chambers or in an oven. It is important that air 
circulate freely during the drying process (Wakely, 1954), and that the tempera- 
ture not exceed 30C. Small lots can be dried chemically with CaO, sulphuric 
acid or CaCl 2 (Magini, 1962). 

Moisture content and storage temperature. For most species of the genera 
Larix, Picea, Pinus, Pseudotsuga, Thuja, Tsuga, Chamacyparis, Cupressus, 
Cryptomeria, AInus , Betula and Eucalyptus , the following can be recommended 
(Schonborn, 1964, Wang, 1974 and Earner, 1975): 

Storage period 

Moisture content % 

Temperature C 

3-5 years 
More than 5 years 
More than 10 years 

6 - 8 % 
6 - 8 % 
6 - 8 % 

+2 to +4C 
-10 to +4C 

In the case of Eucalyptus, Barner (1975) states that most species can be stored 
for a period of up to ten years in airtight containers at temperatures of 1-5 C, 
if the moisture content is kept at 4-8 percent. Turnball (1975), Filho and 
Lisbao (1973) provide more detailed information on some species of Eucalyptus. 

As mentioned above, moisture content and temperature are closely linked. If 
factors fall short of the optimum, it can be offset to a certain extent by the 

Storage containers. Cold storage rooms where the temperature can be kept at the 
desired level are normally very expensive to build, so it is important to store 
the seed in airtight containers. Glass or metal containers were originally 
used for this purpose. More recently, plastic containers and polythene bags 
have begun to be used; however, they are containers not recommended for long- 
term storage, since they are not entirely damp-proof (Owen, 1956). 

Seeds that can be stored 3-5 years with medium moisture content and at low 

a. This group includes Abies (Barner, 1975) and Cedrus and Libocedrus 
(Holmes and Buszewicz, 1958). 

Preparation for storage. Heat should not be used in extracting the seeds. 
The cones should be stored in a well-ventilated place until they begin to 
separate, after which a mechanical handler can be used, very carefully, to 
activate extraction. Care should also be taken in later handling, since 
seeds are extremely sensitive. 

Drying for storage. Abies seeds should not be dried until a 1-2 month 
after ripening period. Drying should be done carefully at temperatures 
not exceeding 25 C. 

Moisture content and storage temperature. 

Temperature C 

Storage period 

Moisture content % 

1-3 years 
More than 3 years 

12-13 % 
7-9 % 

-10 to-20C 

Storage containers. See group 2. 

b. To this group belong the leafy species, Acer, Fagus , Fraxinus, Ulmus, 
and possibly Tectona (Barner, 1975). 

Preparation for storage. With the exception of Ulmus, whose seeds should 
be harvested before maturity, seeds should be fully ripe before harvesting. 

- 66 - 

The seed should always be spread out in well-ventilated rooms and turned 
over at regular interval. 

Drying for storage. Normally it is not necessary or desirable to dry 
seed artificially. 

It should be pointed out that Fagus sylvatica seed should after-ripen 
under cold, humid conditions for another two or three months , after 
which the moisture content can be carefully reduced (Barner, 1975). 

Moisture content and storage temperature. For most species in the group 
the following may be recommended: 

Storage period 

Moisture content % 

Temperature C 

1-2 years 
2-3 years 
More than 3 years 

20 - 25 % 
12 - 20 % 
7 - 10 % 

-4 to -10C 

For Tectona grand is there are no precise data, but it seems that the seed 
can remain viable for 2 to 3 years if it is in dry storage in plastic bags 
(Murthy, 1973). 

Bonner (1978) estimates that the storage period for Triplochiton scleroxylen 
and Gmelina arborea is the same; seed is stored at temperatures of - 5 C 
and moisture contents of 5 - 10 percent. 

Storage containers. See paragrah 2. However, it should be observed that 
seeds with a moisture content of more than 15-20 percent should not be 
stored in airtight containers, but in bags which allow some air to circulate. 

4) Seeds that can be stored 1-3 years with high moisture content and at low tempera^ 

This group includes various leafy species with large seeds and fruits, such as, 
Aesculus, Castanea, Juglans, Liriodendron and Quercus (Barner, 1975). 

The most important principles of storage are good ventilation, uniformly high 
moisture content, and moderate to low temperature. 

Holmes and Buszewicz (1958) provide a series of examples of methods of short-term 
storage (in the winter), both in the open air and in a building or under a 
protective covering. As a general rule, these methods are not used for long-term 
storage; also, it is very difficult to control moisture content and temperature 
during storage. 

5) Very short-lived seeds 

This group includes Salix and Populus (Barner, 1975). 

Normally the seed loses its viability very rapidly, but with some species the 
seed can be preserved longer (Jones, 1962). Holmes and Buszewicz (1958) and 
Wang (1974) have provided more detailed information on individual species* 


The purpose of seed analysis is to obtain actual data on a given seed lot 
(Justice, 1972). 

One of the most important factors in successful testing is the use of standard 
methods providing comparable, replicable results. In this connection, mention 
may be made of the International Seed Testing Association (ISTA), which prepares 
and publishes standardized methods for seed testing. The last updated edition of 
ISTA standards was published in 1976 (ISTA, 1976). 

- 67 - 


For information from a seed lot to be reliable and accurate, the sample taken (and 
this is the purpose of the analysis), should be representative fo the seed lot. It is 
therefore important to use standarized, objective sampling methods. Even very precise 
analyses are less valid if the sample taken is not representative. 

Probes are often used, reaching all parts of the bag and designed in such a way 
that they extract samples of an equal volume. 

Normally, the sample taken should be reduced to a standard portion, and also at this 
point it is important that the reduction should be done objectively and correctly. Turnbull 
(1975) describes various mechanical and non-mechanical methods of reducing the sample. 


Forest seeds can contain impurities in the form of weed seeds, seeds of other tree 

species, bits of seeds, leaves and other things. The purpose of seed analysis is to make a 

quantitative assessment of the composition of the lot under analysis. To this end, the 
lot is divided into its component parts (Turnbull, 1975). 

Pure seed means seeds of one species. As well as ripe seeds, these can be under- 
sized, shrivelled, immature, or germinated seeds, providing they can be identified with 
certainty as belonging to the species in question. 

Parts of seeds more than half the original size (ISTA, 1976) are also included. 
Leguminosae and Coniferae seeds without testae belong to the inert matter group. 

Inert matter includes seedlike structures, for example, broken wings and conifer 
seed coats. 

Foreign elements include sand, stones, leaves, bark and any other element which is 
not seed. 

The entire sample is weighed, including all impurities, after which the pure seed is 
separated out and weighed separately. 

Calculation of the percentage of purity is expressed in the foiling way: 

Purity (7.) = Weight of pure seed 1 

X 1 UU 

Total weight of sample 

Most of the work of dividing the lot to be analysed into the relevant component 
parts has to be done manually, using a magnifying glass, stereomicroscope, etc. However, 
there are today various apparatuses and methods of reducing the manual work to a certain 
extent. First, there is the seed blower, with which it is possible to divide the lot into 
heavy and light componenents. Among other methods may be mentioned: separation by density 
(Stermer, 1964), use of a vibrating table (Guldager, 1973), and X-rays (Kamra, 1965). 

The pure seed component is often used for germination analysis, and is therefore 
important in' any evaluation of the productive potential of a seed lot, provided that purity 
and germination analysis are both tested. 


It is difficult to give a definition of seed viability broad enough to be applica- 
ble in all circumstances. For roost purposes, a seed is considered viable if it germinates 
under favourable conditions, assuming potential dormancy has been eliminated (Roberts, 1972). 

Methods of ascertaining germination potential and methods of indirectly measuring 
seed viability are described overleaf. 

- 68 - 


The fundamental purpose of any laboratory analysis of germination is to estimate the 
maximum number of seeds that would germinate under optimum conditions. 

Germination is expressed as the percentage of pure seed producing normal plants, or 
as the number of seeds germinating per weight unit of the seed lot. It is advisable to 
make germination analyses in accordance with 1STA standards. 

All germination analyses should be made on pure seed. 

The seed should be mixed and counted thoroughly, and the analysis should be repeated 
several times. For more information on details concerning preparation and use of laboratory 
equipment, see Turnbull (1975). 


The choice of germination equipment will depend both on the quantity and the type 
of seed. The most important equipment recommended by ISTA is briefly described below: 

1) Jacobsen and Rodewald apparatus. The seeds are placed directly over a water 
bath or damp sand. The substratum or filter paper bed is kept moist by means of 
a wick from the substratum down to the water. Alternatively, porous dishes 
containing seed can be placed directly over the damp sand or over water. The 
temperature of the container is controlled automatically. 

2) Germination cabinet. In this apparatus trays can be fitted in layers, therefore 
requiring less space. Also, ambient humidity, temperature and light can be 
controlled to a certain extent. 

3) Room Germinator. If a great many analyses are to be made, whole rooms can be 
used with the appropriate equipment to control temperature, humidity and light. 


Optimum conditions for the different stages of germination are not identical and can 
very often vary with different seeds within one seed lot. As a result, most of the research 
on seed has aimed at ascertaining the combination of conditions that provides the most 
uniform, rapid and complete germination of most samples in any one species. 

Turnbull (1975) comments in detail on the following factors which roust be borne in 
mind in germination analyses: 

1) Germination substratum; 

2) Humidity and ventilation; 

3) Temperature control; 

4) Light; 

5) Distance between seeds; 

6) Fungus control. 


The seeds of most tree species germinate upon exposure to the right humidity and 
temperature. However, seeds of some species do not germinate even under favourable condi- 
tions, until they have undergone a physical or physiological change. This state is called 
dormancy. The state of dormancy may be transitory, and many species exhibit dormancy only 
during the first six months after harvesting. 

- 69 - 

Turnbull (1975) mentions the following types of dormancy: 

1) Embryonic dormancy. "Embryonic dormancy 11 means the situation in which germina- 
tion of the fully developed seed appears to be blocked by internal factors. 

2) Dormancy of the testa. Many species of seed do not germinate because they have 
a thick testa. 

3) Induced or secondary dormancy. A few species can enter into a state of dormancy 
due to incorrect storage or handling, and this type of dormancy is called secon- 
dary dormancy. 

4) Unripe embryo. Seeds in which the embryo is not entirely developed when the 
fruit is ripe require a period of after-ripening before germination. 

5) Mechanical resistance of the testa. In some cases considerable force is needed 

to break the testa. This applies to certain species of Prunus and Tectona grandis. 
As in the case of dormancy of the testa, scarification can be used to break the 

6) Double dormancy. Seeds of some species have two types of dormancy that have to 
be eliminated before they germinate. The most usual combination is embryonic 
dormancy and testa dormancy. 


A seed is said to have germinated when it has developed into a normal plant. 
Various groups of abnormal plants (see more details in ISTA, 1976) are not included in 
germination trials because they seldom survive. Germination tests (even excluding ab- 
normal plants) normally produce over estimates of viability under field conditions, due 
to the optimal testing conditions. 

The purpose of quick viability trials (Turnbull, 1975) is: 

1) To ascertain quickly the viability of species that usually germinate slowly 
and show signs of dormancy in normal methods of germination; 

2) To ascertain the viability of lots which, at the end of the germination test 
show a high percentage of hard or ungerminated fresh seeds. 

Four different methods are described below: 

1) Analysis by cutting open. By opening the seed it is possible to study the 
endosperm and the embryo directly and to see whether their colour is normal 
and if they are developing normally. The method is not very accurate and 
often leads to over estimates of germinative capacity. 

2) Tetrazolium tests. By this method, living cells turn red when the colourless 
tetrazolium salt is reduced. The procedure has been described in detail by 
ISTA (1976). 

Justice (1972) points out that the use of this method is in practice limited 
by a series of problems, such as the difficulties in staining some seeds; 
some seeds have to be opened for the colour to become visible. In several 
cases, results are not in line with those of germination tests, and inter- 
pretation of the different shades is difficult. 

3) Trials of extracted embryos. Extracted embryos are placed on damp filter 
paper in Petri dishes. In a few days it is possible to see which embryos are 
alive and which are not. 

- 70 - 

4) X-ray technique. This method reveals empty seeds, mechanical damage, abnormal 
development of the internal structure of the seeds, and the thickness of the 
testa. It also assesses viability, using a technique of staining or contrasting 

The results of the viability tests may be described as follows (Turnbull, 1975): 

1) Percentage of germination. This is the percentage of seeds in the lot which 
have germinated by the end of the trial. 

2) Germi native potential. This is the sum of the germinated seeds and the remaining 
sound ungerminated seeds. 

3) Germination energy. This indicates the percentage germinated within a given 
time, for example, 70 percent within seven days, 90 percent within 14 days, etc. 

4) Number of live seeds per weight unit. The indication could be useful to the 
people who are going to use the seed. 


As mentioned above, determination of moisture content is very important for storage 

1STA recommends the following three methods of determinating moisture content (1STA, 

1) Oven drying method (130 C). Owing to the high temperature, this method cannot, 
however, be used for forest seeds. 

2) Oven drying method (105 C). 

3) Toluene distillation method. This method should be used to determine the moisture 
content in seeds containing volatile oils, for example, Abies spp. 

As well as the methods mentioned above, there is a series of automatic measurements 
of moisture content. They are not sufficiently accurate for official trials, but they can 
provide rapid and fairly reliable information on the moisture content of seeds. 


Determining weight of 1 000 seeds has been described by 1STA (1976). 


Health tests are run to determine the presence of micro-organisms or diseases in 
seeds. These trials are important for three reasons: 

1) Disease-carrying seeds can lead to forest outbreaks of pests, lowering the 
commercial value of the trees. 

2) Through seed lots, diseases can spread to new regions. Quarantine investigation 
and certification for the international trade may therefore be necessary. 

3) Health tests provide more data on the causes of abnormal seddlings, and could 
supplement germination analysis. 

The 1STA standards provide general rules for health tests. 

- 71 - 


Both the seed testing equipment and its use are described by ISTA (1976). 


The purpose of certification of forest seeds and plants is to preserve and make 
available to working foresters sources of seeds, plants and other propagating material 
of superior provenances and cultivars that have been cultivated and distributed, thus 
ensuring the genetic identity and superior quality of seeds and plants (Matthews, 1964). 


Barber (1969) summarized the problems relating to control of genetic identity in the 
following way: 

Good results in tree improvement and afforestation programmes depend on accurate 
control of the genetic identity of reproductive material. We should try to use only 
reproductive material of known genetic identity. However, the precision with which we can 
identify the material will vary according to species, locality and final use. 

The forest tree breeder must have full knowledge of the source of the seed plasm 
with which he is working. Tt is particularly important that the identity of each tree be 
maintained, so that the researcher may consider the risks of any negative trait from the 
mating of related individuals. As offspring are produced and grow, the breeder needs to be 
able to study the seneolofcical record of each individual to pinpoint which parent has 
contributed which character, whether desirable or not. He/she should be in a position to 
duplicate all crossings if necessary, and must catalogue and identify accurately all the 
material exchanged or approved for use. 

The forester should know which source or line best meets his needs. To achieve 
optimal results, he/she should know the exact source of the material used for the estab- 
lishment of plantations and for stand regeneration; knowledge of the genetic identity 
of the material used is necessary to plan proper spacing and cultivation techniques. As 
the use of a disease-resistant strain, for example, lowers mortality and produces fewer 
defects, the forester can space trees further apart or thin more frequently. Reports of 
major genotypic interaction in the environment indicate that tree breeders can develop 
cultivars that react favourably to qualitative difference in site or cultivation methods. 
If seeds or seedlings are not available from the appropriate source, it could be desirable, 
from the economic point of view, to postpone planting for a year or more. 

For a certification programme to function satisfactorily, there should be a mechanism 
to facilitate the participation of all interested parties in the formulation of operational 
procedures. Certification should be backed by legislation. 

Several countries use certification programmes (Matthews, 1964). Some programmes, 
such as the North American programme, are regional. 

An international programme under OECD was established in 1967. In 1974 a few 
modified rules were prepared (OECD, 1974). 

The OECD programme in based on the voluntary participation of member countries, but 
NATO member countries can also participate. 

A broad programme should include the following elements: 

1) Preparation of maps, with an indication of the distribution of important species. 

2) Delimitation of regions of provenance of these species. 

- 72 - 

3) Delimitation of important regions for afforestation and reforestation. 

4) Estimate of supply and demand of seeds and plants. 

5) Organization and administration. 

6) Classification and approval of sources. 

7) Recommendations for the choice of provenances and transfer of reproductive 

8) Production and control measures. 

9) Record of data and documentation. 
10) Sale of reproductive material. 

Ref. items 1-A. These items are important for evaluation of the need to establish a 
certification programme. An estimate of the quantity of seeds and plants necessary should 
be made and compared with available sources, both on the site and elsewhere. 

Ref. item 5. This item includes the designation of the various groups of authorities 
responsible, such as: 

1) a director or directors; 

2) an advisory group; 

3) a working group for approval of sources; 

4) inspectors. 

Ref. item 6. This item is fundamental to the certification programme, and has been 
discussed in detail by Barner (1973). Major sources could be listed as follows: 

1) Regions of provenance; 

2) Stands; 

3) Regions of seed production; 

4) Individual trees; 

5) Seed plots. 

Approval of sources, in accordance with certain minimum requirements, would be the respon- 
sibility of the above-mentioned working group for approval of sources. 

A national list of sources should be prepared. The list should include the most important 
data, such as Latin name, identification, location, origin, ecological conditions and 
results of trials. A definition of classes and instructions for collection of information 
has been provided by Barner (1975). It is not necessary that the national list include all 
the above-mentioned types of source, but it should be possible to extend the programme to 
include all types later. 

Ref* item 7. One later development of the approval of sources could be the recommendation 
of these sources for general or specific uses. Sources have often been approved on the 
basis of the phenotypical appearance of the seed trees or on the basis of very limited 
progeny texts. Recommendations should be based on tests, but since these take a long time 
and seed must be produced, it is necessary to use sources of seed that have not yet been 
tested. In certification programmes, therefore, a distinction is made between tested 
material and untested material within each class of the approved sources. The certification 

- 73 - 

programme should be established and developed in close collaboration with forestry research 
and particularly with forest tree breeding. 

Ref . item 8. As an example of rules relating to production and control measures, Barner 
(1975) indicates the minimum standards and requirements of the OECD programme as regards 
material from approved sources. The standards establish how inspection is to be carried 
out in the forest, the plantation and the nursery, and how to mark and package seeds. 

Ref. item 9. Recording data and documentation is an integral part of the control procedures. 
Each seed lot, for example, should be accompanied by standard data with detailed indications 
of identification, harvesting, seed extraction storage and quality. 

Ref. item 10. Once all the operations have been carried out in accordance with the programme, 
a certificate can be granted, to which reference can be made when material from the source 
in question is sold. 


Barber, J., 1969. Control of Genetic Identify of Forest Reproductive Material. Second 
World Consult, on For. Tree Breeding, Washington, D.C. Vol. 2, 11/3. 

Barner, H., 1973, Classification of Sources for Procurement of Forest Reproductive 
Material. FAO/DANIDA Training Course on Forest Tree Improvment. Limuru, Kenya. 

Barner, H., 1975. The Storage of Tree Seed, FAO/DANIDA Training Course on Forest Seed 
Collection and Handling, Chiang Mai, Thailand, FAO 1975. 

Barton, L.V., 1961. Seed Preservation and Longevity. Leonard Hill, London. 

Bonner, F.T., 1978. Storage of Hardwood Seeds, Forest Genetic Resources Information 
no. 7, Forestry Occasional paper 1978/1, FAO, Rome. 

Christensen, C.M., 1972. Microflora and seed deterioration. In: Viability of seeds by 
E.H. Roberts. Syracuse University Press. 

Ezumah, B.S., 1976. Seed Handling and Storage, Savanna Afforestation in Africa, Kaduna, 
Nigeria, FAO 1976. 

FAO, 1955. Handling forest tree seed. FAO, Rome 

Filho, W.S. & Lisbao, L., 1973. Influence of the relative humidity on the characteristics 
of the seeds of Eucalyptus saligna. IUFRO Internat. Symposium on Seed Processing, Bergen 
Norway 1973. Publ. The" Royal College of Forestry, Stockholm. 

Frankel, O.H., 1970. Genetic Conservation in Perspective, Genetic Resources in Plants - 
their Exploration and Conservation, 1970. 

Guldager, P., 1973. Seed problems related to direct sowing in pots, "Seed Processing" 
Proc. IUFRO WKG. Group on Seed Problems, Bergen. Vol.11, Paper 12. 

Harrington, I.F., 1970. Seed and Pollen Storage for Conservation of Plant Gene Resources, 
Genetic Resources in Plants - Their Exploration and Conservation. 1970. 

Holmes, G.D. & Buszewicz, G., 1958. The storage of seed of temperate forest tree species. 
Forestry Abstracts, Leading Article, for. Abs. Vol. 19. Nos. 3 and 4. 

ISTA, 1976. International Rules for Seed Testing, Rules 1976, Seed Science & Technology 4. 

Jones, LeRoy, 1962. Recommendations for successful storage of tree seed. Tree Planters 
Notes. For. Service U.S. Dept. of Agric. No. 55. 

Justice, O.L, , 1972. Essentials of seed testing, "Seed Biology", III, (Ed. T.T.Kozlowski). 
Academic Press Inc. New York. 361-370. 

- 74 - 

Kamra, S.K.; 1965. The use of X-ray radiography for studying seed quality in grasses. 
Proc. Int. Seed Test. Ass. 30(3): 519-524. 

Kamra, S.K., 1967. Detection of mechanical damage and internal insects in seed by X-ray 
radiography. Svensk Bot. Tidskr. 61, Stockholm. 

Magini, ., 1962. Forest seed handling, equipment and procedures. Unasylva, Vol. 16, 
FAO Rome. 

Matthews, J.D., 1964. Seed Production and Seed Certification. Unasyla, Vol. 18. (2-3). 

Moore, R.P., 1969. History supporting tetra2olium seed testing. Proc. Int. Seed Test. 
Ass. 34. 

Moore, R.P., 1972. Effects of mechanical injuries on viability. Viability of seeds by 
E.H. Roberts. Syracuse University Press. 

Murthy, A.V.R.G. 1973. Krishna. Problems of teak seed. 1. Flower and fruit studies. 
2. Germination studies. Intern. Symposium on Seed Processing, Vol. II, Bergen 1973. 
The Royal Coll. of For., Stockholm. 

OECD, 1974. OECD Scheme for the Control of Forest Reproductive Material Moving in Inter- 
national Trade. OECD, Paris. 

Owen, E. Biasutti, 1956. The storage of seeds for maintenance of viability. Commonwealth 
Agri . Bureaux, Bull. No. 43, Farnham, England. 

Roberts, E.H., 1972. Viability of seeds. (Ed. E.H. Roberts), Chapman and Hall, London. 
448 pp. 

Schonborn, A. von., 1964. Die Aufbewarung des Saatgutes dar Waldbaume. Bayrischer 
Landwirtschaftsverlag, Munchen. W. Germany. 

Stermer, R.A., 1964. Purity analyses of certain grass seeds by flotation techniques. 
Proc. Ass. Off. Seed Anal. 54: 73-81. 

Turnbull, J.W., 1975. The handling and storage of eucalypt seed. FAO/DANIDA Training 
Course on Forest Seed Collection and Handling. Chiang Mai, Thailand, FAO, 1975. 

Wakeley, P.C., 1954. Planting the Southern Pines. Forest Service. U.S. Dept. of Agric. 
Washington D.C. 

Wang, B.S.P., 1974. Tree seed storage. Dept. of Environment, Can. For. Service, Pub. 
No. 1335, Ottawa. 

Wang, B.S.P., 1977. Procurement, Handling and Storage of Tree Seed for Genetic Research. 
Third World Consultation on Forest Tree Breeding, Canberra, Australia 1977. 

- 75 - 


B. Ditlevsen 
National Forestry Service, Denmark 



Introduction 76 

Requirements of a good experiment 76 

No systematic errors 76 

Precision 76 

Field of application 77 

Simplicity 77 

Estimate of uncertainty factor , 77 

Principles of experimental designs 78 

Null hypothesis 78 

Purpose of the test 78 

Experimental errors 78 

Repetitions 78 

Randomization 79 

Local control 79 

Considerations of a practical nature 80 

Designs 80 

Complete randomization 80 

Random block design 8l 

Latin square design 82 

Incomplete block design 83 

Split plot design 8U 

Bibliography 85 

- 76 - 


The principles governing experimental designs may be expressed in the following way 
(Brown et. al. , 1977): 

1) Observation of a phenomenon. 

2) Formulation of a hypothesis on the phenomenon. 

3) Testing the hypothesis. 
h) Application of Findings. 

Tests are tools to test established hypotheses. 

Successful test design and implementation are most important factors underlying 
the soundness of test conclusions (LeClerg, 1967)- 

Forest tree improvement tests have the following main objectives: 

1) To evaluate the material for genetic improvement with respect to a series of 
desired properties, based on an established hypothesis. In most cases, these 
are field trials, where the material is tested under conditions similar to the 
growing conditions (habitat conditions) of the improved material. However, 
they could also be nursery, greenhouse or laboratory trials. 

2) To study the fundamental genetic parameters which are important for research 
and for future development in the field of forest tree improvement. 

3) TO analyse different treatments, methods of genetic improvement, etc. Like the 
studies mentioned under point 2, these analyses should be made initially with a 
view to further research in this field. 

Irrespective of the purpose of the test, all the different practical factors should 
be taken into account in planning it, so that maximum information is obtained. 


Cox (1958) has established the following five considerations in planning a test. 

No systematic errors 

This means that experimental units containing a particular material, for example a 
particular provenance, should not deviate in any systematic way from the other provenances 

in the test. 


If there are no systematic errors, the probable magnitude of random error in the 
estimate of the effect of the treatment (material), can generally be measured through 
standard error. Standard error (SE) is calculated as follows: 

(x-x) 2 s 

SE ; 

n-1 VT 

B = standard deviations X * individual observation 

X = the estimated mean of X values n * number of observations 

- 77 - 

The precision of a test will depend on the following factors: 

a. The real variability of the experimental material and the accuracy with which the 
experimental work has been done. 

b. The number of experimental units (and the number of observations repeated per 

c. Experimental design (and the method of analysis, if the latter is not completely 

A general requisite of precision is that the standard error be sufficiently small to 
allow us to draw convincing conclusions; on the other hand, it should not be too small. If 
the standard error is great, the test will be almost valueless, whereas an unnecessarily 
small standard error implies a loss of experimental material. 

Field of application 

Estimating the difference between two materials in a test, we reach conclusions that 
refer to the special set of units used and to the conditions governing the test. If we 
wish to apply the results of the test to other conditions or units, another uncertainty is 
added, as well as the uncertainty measurable as standard error. It is important that experi- 
mental material be tested under conditions not essentially different from those of the 
sites where it is planned to use the material. 


This is a very important factor which should always be considered in planning a test. 
By simplicity is meant not only a simple design, but also that it is desirable to use simple 
methods of analysis. Fortunately, in most cases, efficiency in design is matched by simpli- 
city in methods of analysis. 

Estimate of uncertainty factor 

It is desirable to estimate the uncertainty factor of the assumed differences between 
materials. This usually means estimating the standard error for these differences, and then, 
on this basis, calculating the limits of error (limits of reliability) for real differences 
at a given level of probability. From this, the statistical significance of the difference 
between two materials can be measured. 

The standard error of the difference is calculated as follows: 

SE - 

n. and n 2 s number of observations on which the two estimated means are based. 
The limits of reliability for real differences may be calculated as follows: 

XI and X2 the two estimated averages 

t value of student's t at a given level of probability 

n. and n 2 * number of observations on which XI and X2 are based. 

- 78 - 


Series of fundamental and general factors in experimental designs is described 

Null hypothesis 

As mentioned in the introduction, the purpose of the tests is to test a hypothesis 
based on existing differences between materials or treatments. Generally, a so-called null 
hypothesis is established, which assumes that any possible differences can only be attribu- 
ted to contingencies. This null hypothesis is used in combination with a test of signifi- 
cance as an alternative to the established hypothesis being studied. 

The concept of null hypothesis is very important in experimental designs. When we 
set up the test, we know some factors that produce differences between the units of the 
test, but there will always be differences not attibutable to specific factors. To use the 
null hypothesis as an alternative hypothesis, there must be random distribution of causes 
of variation that cannot be attributed to specific factors. 

Purpose of the test 

It is important to define the purpose of the test. For example, the purpose could 
be the study of a series of consignments of seed for subsequent selection of the best con- 
signments of seed for further genetic improvement. In many situations the test are multi- 
factorial, and in such situations it is essential that the purpose be formulated before the 
design is chosen. For example, it would be desirable to know whether the purpose is the 
study of one or more of the principal factors of the test, or whether it is rather an 
estimate of a specific interaction. 

It is also important to formulate the field of validity to which the results will 
be applied. 

Experimental errors 

Experimental errors were described by Fischer (1951) as the lack of uniform gains 
in plots which had, on the whole, received the same treatment. The experimental errors that 
occur in nursery and field tests are, in the first place, caused by differences in soil 
conditions. Previously attempts were made to reduce this error by holding tests under 
homogeneous soil conditions using an improved experimental technique. However, experimen- 
tal error cannot be completely eliminated, not even under apparently identical conditions. 
Instead, attempts are made to measure the error, so that the uncertainly affecting the 
conclusions drawn from the data material can be estimated. This then determines the 
selection of an experimental design in a given situation. A good experiment is described 
by LeClerg (1967) as an experiment that estimates the magnitude of uncontrolled experi- 
mental error. 

Fischer (1931) formulated the following three decisive principles for obtaining 
exact data - principles which form the basis of the experimental design to be established 
and of the subsequent statistical analysis: 

Repetitions* There are usually two possible ways of reducing error in the estimate 
of an effect. One consists in improving the experimental technique, for example, by using 
more precise measuring instruments and taking greater care in the collection of data, etc. 
The other possibility and, in many cases, the only one practicable, is to repeat the test 
a number of times and use the average of the results obtained. Unfortunately, the second 
method is not very efficient, since the error of the average is only reduced by the square 
of the number of repetitions. Thus the average of four repetitions will only reduce the 
error of one-half ( |/7J of the error of the individual repetition. As well as reducing 
the error of the estimate, repetitions make it possible to estimate the uncertainty factor 
in the experimental results, e.g. calculating intervals of reliability. Repeated measure- 
ments in the same plot are not considered as repetitions but can only be considered as 
sub-samples, and the variation between them constitutes the sampling error. 

- 79 - 

Randomization. In any experimental design it is very important, according to Fischer 
(1926), that the different units of the experiment should be randomized. Fischer (1926) 
considers that "the estimate 11 of experimental error obtained through repetitions depends 
on existing differences between plots treated in the same way, An estimate of experimen- 
tal error will only be valid for its purpose if, in the arrangement of the plots, we en- 
sure that pairs of plots that have received the same treatment are neither closer to- 
gether nor further apart nor in any other related way different from plots that have re- 
ceived different treatments". 

In most cases, randomization is applied to the experimental area; however, it can 
also be applied to time. 

In practice, randomization can be done in several different ways. In tests of limited 
scope, it can, for example, be done by consecutively numbered cards, which are shuffled and 
picked at random. Depending on the design in question, randomization can apply to the complete 
test or separately to individual lines, columns, etc. This is described in greater detail 
for individual designs. If the tests are broader in scope, randomization tables (or, where 
a computer is available, standard randomization programmes) can be used. 

Local control. Local control means certain restrictions on the random distribution 
of treatments or materials in a repeated test, so that the part of total variation not 
applicable to comparisons of treatments or materials may be eliminated. Local control 
includes the arrangements of blocks and plots in the test. By arranging the blocks in such 
a way that the variation of, for example, soil conditions within the blocks, is minimized, 
while a considerable variation can be found between the blocks, an important improvement 
will be achieved in the precision of the test, since the variation between blocks can be 
calculated and eliminated when the data is analyzed. 

Plot size and shape, plant spacing, etc., can often vary from one place to another, 
depending on the experimental area, the purpose of the test, the volume and quality of the 
material, and on other factors often of a practical nature. Concerning plot size and shape, 
there is a series of studies which illustrate optimal plot size and shape (Johnstone and 
Samuel, 1974). It does not seem possible to set any general standards for the arrangement 
of the plots, but the following points are important when choosing both plot size and shape: 

1) An increase in plot size will produce a more reliable estimate of the plot mean. 
On the other hand, an increase in plot size will lead to a corresponding increase 
in the size of the blocks and therefore in the variation within the individual 
block. A greater variation within the blocks will reduce the possibilities 

of discovering intraplot differences, if any. Thus Wright and Freeland (1959, 
1960) observed that a reduction in plot size to single tree plots led to more 
effective tests on pine height and diameter. 

2) For tests of, for example, different fertilizers, fairly large plots must be 
used, perhaps surrounded by buffer strips to avoid any influence from adjacent 
parcels. The same applies to tests in which, with time, there could be 
competition between trees of adjacent plots. 

3) As regards plot shape, in most cases it will be better to select square plots, 
thus obtaining the smallest perimeter and therefore minimal influence from 
adjacent parcels. However, where there are considerable systematic site 
variations, such as sloping land, it may be better to select oblong plots 
arranged in such a way that this variation is absorbed within the plots. 

As mentioned above, the precision of a test can be improved primarily by increasing 
the number of repetitions. On the other hand, a useless increase of precision through a 
large number of repetitions will lead to a loss of resources in the form of plants and 
labour. It is therefore important to estimate, when preparing the test, the number of 
repetitions needed to give the required precision. For example, if it is desired that 
a 10 percent difference of materials show as significant at a level of significance of, 
for example, 5 percent, the necessary number of repetitions could be calculated via the 

- 80 - 

following formula: 


LSD = real difference from D 

t = Student's t at the desired level of significance (5 percent) and with the same number 
of degrees of freedom as the experimental error (S) 

S * experimental error 
n number of repetitions. 

If the approximate magnitude of the experimental error is known, either from previous 
tests of the same kind, or on the basis of a general knowledge of the conditions, the 
equation could be solved with respect to n, as follows: 

t 2 2 S 2 
na t . 2 . S 

Consideration of a practical nature 

As well as considerations of a theoretical nature with regard to a test, there are 
often practical constraints setting up the test. 

In the first place, the area available can lead to a series of limitations, which 
will oblige the planner to omit several blocks, repetitions or treatments, or to use a 
different and often less efficient design. The choice of one or another possibility will 
depend on factors related both to the experimental material and to the site. If, for 
example, there is a very small variation in soil conditions, the number of repetitions 
will naturally be reduced. Another possibility is to reduce the number of treatments or 
materials, if it is known beforehand that some are homogeneous and can perhaps be lumped 
together in the analyses. 

In the second place, there may be problems in setting up the test at several different 
sites. If it is desired to apply the results more broadly, it will be necessary to set up 
tests at a number of sites representative of the field as a whole. 

As well as these factors, it should be borne in mind that some parts of the test 
could be damaged or destroyed. If, for example, there is any risk of damage, a relatively 
sound design should be selected that could yield valuable information, even if some 
values are lacking. 

Similarly, it is important that both the design and the methods of analysis be rela- 
tively simple, particularly when the people conducting the experimental work are not fully 
conversant with more complicated methods. 


With time, various different experimental models have been developed, some based on 
the principle of random distribution, others on that of automatic distribution of treatments 
or materials in the plots. A series of models is described below, which largely meet the 
requirements of a good experiment as described in the preceding section (Jeffers, 1960, 
Perarce, 1975). 

Complete randomization 

In this design there is no grouping of experimental units, and the different treat- 
ments are assigned to individual plots completely at random. 

- 81 - 

The method is very simple and flexible and can be used to advantage under the follow- 
ing conditions (Cox, 1958): 

1) In very small tests where it is important to have the maximum number of degrees 
of freedom in the estimation of error. In a test which contains N plots and a 
treatments, there are N-a degrees of freedom, whereas in a test with an equal 
distribution of blocks, there will be N-a-b+1, where b is equal to the number 
of blocks in the test. 

2) In tests where, as far as can be seen, there is no immediate reason for establish- 
ing a block design. This could be the case, for example, in laboratory tests 
where the environmental conditions are controlled and uniform. 

3) In tests where an increase in precision can only be achieved through the use of a 

An example of a complete randomization design is illustrated in Figure 1. 




















Figure 1. Complete randomization design with three treatments (A,B,C), each one 
repeated six tiroes. 

Random block design 

In cases where it is hoped that there may be some form of variation in environmental 
conditions, complete randomization is not very practical, since the variation between plots 
treated in the same way, and therefore the experimental error, will be large, and precision 
considerably reduced. 

Under such conditions, plots are usually grouped in blocks, so that each block 
includes all the different treatments once. With this design, variation between blocks 
can be eliminated by statistical analysis, reducing the experimental error and increasing 
the precision of the test. Random block designs are simple to establish, and Figure 2 
shows an. example of a test that includes four treatments and four blocks. The distribution 
of the treatments within each individual block must be random. 

Block 1 

- 2 

- 3 

















Random block design with four treatments (A,B 9 C,D) and four blocks. 

As mentioned above in connection with local control, blocks should be arranged in 
such a manner as to absorb the maximum variation between blocks, while variation should 
be maintained within blocks at as low a level as possible. Figure 3 shows examples of a 
good arrangement of blocks under the following circumstances: 

- 82 - 

1* Establishment in an area with a systemetic environmental variation, such as un- 
even ground^ near a window in a laboratory, etc. 

2. Establishment in an area where there is marked variation in soil conditions; the 
plots should be adjoining, irrespective of their physical arrangement, in blocks 
within which conditions appear uniform. 

Block 1 

Block 2 
Block 3 
Block 4 


Block 3 

Figure 3. Arrangement of blocks in a random block design under conditions of 
1) systematic site variation and, 2) considerable site variation. 

The random block design is simple to establish, and analysis and interpretation 
of the test are also simple. At the same time, this type of test is very sound, since 
the analysis can be made, even if some values are missing, without any major problems 
(Brown et al., 1977). 

In principle, this design can be used for testing a random number of treatments, 
but in practice problems arise to the extent to which the size of the blocks and the 
number of treatments are increased. With the increase in size of the blocks, the varia- 
tion within the blocks is also increased, and the purpose of establishing blocks cannot 
be satisfactorily achieved. In such cases it may be decided to use an incomplete design 
(see below) or a reduction in the number of trees per plot. 

Latin square design 

In this design, plots have been double-grouped, dividing the test into lines and 
columns. A specific treatment is made once in each line and column, as illustrated in 
Figure 4, From this it may be concluded that the number of plots in the test will be 
equal to the square of the number of treatments. 

This design is very effective where there are variations in two directions (for 
example in greenhouses), but at the same time there are some restrictions with respect 
to the number of different treatments that can be studied. For example, ten will require 
the establishment of 100 parcels, and four or less will give too few degrees of freedom 

- 83 - 

to enable a valid test of the treatment to be made. 

In this design there are certain restrictions on the possibilities of randomization, 
but partial randomization can be used, as described below. 

1) One of the formal designs in Fisher and Yates (1957) is chosen at random. 

2) The columns are arranged at random. 

3) The treatments are designated at random by letters A, B etc., in the formal 




























Figure 4. Latin square design with five treatments (A,B,C,D,E). 

Incomplete block design 

These designs are called incomplete designs; that is to say, the number of treatments 
within a block is less than the total number of treatments which are being tested. 

In lattice designs the incomplete blocks are grouped so that each group forms a 
complete repetition; see Figure 5. 

Rep. x 
XI X2 X3 X4 

Rep. Y 
Yl Y2 Y3 Y4 

Rep. Z 
Zl Z2 Z3 Z4 





































Figure 5. Rectangular lattice, 3x4, with three repetitions (X,Y,Z). 

The number of treatments in a lattice design should be the same as the square 
(4x4= 16,' 5 x 5 = 25 etc.) or the product, according to the following formula: 
k(k+l) (3 x 4 = 12, 4 x 5 = 20, etc.). These two types are called square lattice design 
and rectangular lattice design, respectively. In each case k plots (3,4,5, etc*) are 
arranged within each block and k square) or (k+1) (rectangular) blocks within each 
repetition. A completely balanced square design will require k+1 repetitions. 

It is not possible to establish completely balanced designs for k2 = 36, 100 or 144, 
for rectangular designs. 

In establishing a lattice design randomization will follow these rules: 

1) Use as point of departure a formal design, as illustrated, for example, by 
Cochran and Cox (1957). Choose at random the number of repetitions desired 

2) Random lie the order of succession of the repetitions* 

3) Randomize the order of succession of the incomplete blocks within each repetition. 

4) Randomize the plots within each block. 

5) Assign at random different treatments to the numbers of treatment in the formal 

Experimental layouts in the field have the same requirements as those described for 
random block designs, i.e., that the blocks should be so placed that a variation in the 
experimental area is absorbed as an inter-block variation, while the individual blocks are 
as homogeneous as possible. 

The lattice design is useful, and at least as exact as a random block design, with 
the same number of repetitions. However, the statistical calculations involved are 
relatively complicated; it could be difficult, especially in the case of missing values, 
to make a satisfactory analysis. 

Nevertheless, the lattice design has the great advantage that it can always be 
analyzed and interpreted as an ordinary random block design. 

Split plot design 

A split plot design is a special factorial design, inasmuch as it includes two 
degrees of repetitions. A set of factors is linked to the larger plots, while the other 
set or sets of factors are linked to plots within each one of the larger plots. In other 
words, the larger plots are sub-divided into a number of sub-plots; see Figure, 6. 

As there are fewer plots than sub-plots, the factors linked to the larger plots and 
the sub-plots respectively will be tested with different degrees of precision. It is 
therefore an advantage to use this design in situations in which one of the sets of 
factors is of special interest, or in situations in which a specific treatment will 
require a relatively large area (for example, in fertilizer trials). 

The larger plots can be arranged, for example, in random block designs, Latin square 
designs or corresponding designs. The allocation of sub-parcels must be random. 

Rep.l Rep. 2 Rep. 3 Hep. 4 

Figure 6 Design of split plots with two larger plots within each repetition 

(indicated by shading or no shading) and three sub-plots within each larger 

- 85 - 


Brown, A.G. & Ma the son, A.C., 1977. Statistics: Design of Experiments. International 
Training Course in Forest Tree Breeding, Canberra, Australia. 

Cochran, W.G. & Cox, G.M., 1957. Experimental Designs. John Wiley & Sons, Inc., New York, 

Cox, D.R., 1958. Planning of Experiments. John Wiley & Sons, Inc., New York. 
Fischer, R.A. , 1926. The Arrangement of Field Experiments. I. Min. Agr. England 33. 

Fischer, R.A., 1931. Principles of Plot Experimentation in Relation to the Statistical 
Interpretation of the Results, Rothamsted Conference XIII. 

Fischer, R.A., 1951. The Design of Experiments, Hafner Publ. Co., New York. 

Fischer, R.A. & Yates, F., 1957. Statistical Tables for Biological Agricultural and 
Medical Research, Oliver & Boyd. 

Jeffers, I.N.R., 1960. Experimental Design and Analysis in Forest Research. Almquist & 
Wiksell, Stockholm. 

Johnstone, R.G.B. & Samuel, C.J.A., 1974. Experimental Design for Forest Tree Progeny 
Tests with particular Reference to Plot Size and Shape, Proceedings IUFRO Joint Meeting, 

LeClerg, E.L., 1967. Significance of Experimental Design in Plant Breeding, Plant Breeding, 
R.J. Frey (ed), The Iowa State University Press, Ames, Iowa, USA 

Pearce, S.C., 1975. Field Experimentation with Fruit Trees and other Perennial Plants, 
Commonwealth Agricultural Bureaux. 

Wright, J.W. & Freeland, F.D., 1959. Plot Size in Forest Genetics Research, Pap. Mich. 
Acad. Sci., Arts and Letters 44. 

Wright, J.W. & Freeland, F.D., 1960. Plot Size and Experimental Efficiency in Forest 
Genetics. Research, Mich. State Univ. Agric. Exp. Sta. Techn. Bull. No. 280. 

- 86 - 


B. Ditlevsen 
National Forestry Service, Denmark 



introduction . . o / 

Fixed and random effect models ... ... . .. . .. . 87 

Statistical significance ... ... ... ... ... ... ... ... 87 

Tests ... ... ... ... ... ... ... ... tJo 

Cni s Qua re test . . ... ... ... ... ... ... ... 07 

t test ... ... ... ... ... ... ... 07 

f test ... ... ... * ... ... ... ov 

Preliminary investigation ... ... ... ... ... ... ... ... 90 

Outliers and missing values ... ... ... ... ... ... ... 90 

Tests of requisites ... ... ... ... ... ... ... 90 

Normality test ... ... ... ... ... ... ... ... ... 90 

Homogeneity of variance test .... ... ... ... ... ... ... 91 

Additivity test ... ... ... ... ... ... ... ... 91 

Transformation ... ... ... ... ... ... ... ... ... 91 

Contingency tables ... ... ... ... ... ... ... ... ... 91 

Variance analysis ... ... ... ... ... ... ... ... ... 92 

Complete randomization ... ... ... ... ... ... ... ... 93 

Random block, design .. ... ... ... ... ... ... ... ... 94 

Latin square design .. ... ... ... ... ... ... ... ... 95 

Incomplete block design ... ... ... ... .. ... ... 96 

Split plot design ... ... ... ... ... ... ... ... ... 97 

Correlation analysis ... ... ... ... ... ... ... ... ... 98 

Regression analysis ... ... ... ... ... ... ... ... ... 98 

Covariance analysis ... ... ... ... ... ... ... ... ... 99 

Variability components . ... ... ... ... ... ... ... ... 100 

Bibliography ... ... ... ... ... ... ... ... ... ... 102 

- 87 - 


It is important when preparing a test to define the answers one hopes the test will 
provide, and the methods of analysis one plans to apply in appraising the results of the 
test. Thus, the possibilities of drawing sound conclusions from a test will depend to a 
considerable extent on whether or not the test is properly planned. In other words, one 
requisite for a good test will be clear formulation of the purpose of the test, the use of 
a design that can give the required answers, and careful control of the preparation of 
the test. 

Using statistical analysis methods, tree breeder can make a statistical interpre- 
tation of the test. A knowledge of the experimental design of the material and the 
treatments sampled, as well as of the statistical results, are usually sufficient basis 
for a later statistical interpretation of the results. 

Some general aspects of statistical interpretation are described below, giving 
examples of the most widely used statistical methods. Genetic interpretation will be 
mentioned only briefly, since this question is discussed in greater detail in another 


In making an analysis of tests, it is essential to define the statistical model to 
be used. A distinction is made between so-called "random effect" and "fixed effect" part 
(sample) chosen at random from the total of possible treatments. The test results are 
applicable to all treatments represented by the treatments sampled. 

In a fixed effects model, the purpose is to study a specific set of treatments in 
such a way that the test results are not applicable to a group of treatments greater than 
those in the test. 

As well as the models mentioned above, there are also mixed models, in which one or 
more treatments are considered random, while others are considered fixed. 

It is important to define the type of model to be used in each individual case, 
since both the test of the treatments and the computing of the variability components 
will depend on this. Also, the final conclusions and the scope of the test will naturally 
depend on these basic considerations. 

However, very often it can be difficult to decide if the effects of a test should be 
considered random or fixed, and it is difficult to establish univocal standards in this 
respect. The problems have been thoroughly discussed by Kempt home (1975), who also gives 
examples of how the two different approaches influence the variance analysis test. 


The point of departure for the concept of significance is the establishment of a 
null hypothesis, and possibly of an alternative hypothesis, on the test material. Roughly 
speaking, the null hypothesis is that hypothesis which, as problems appear, it is neither 
natural nor advisable to reject, unless observations indicate very clearly that there are 
other possibilities. 

Significance tests are used to decide whether results obtained from a test are in 
conformity with the established hypothesis. These are tests to check whether the results 
deviate significantly from the results anticipated under the null hypothesis. 

A test size is calculated on the basis of the trial and its theoretical distri- 
bution under the null hypothesis is known. If the calculated size of the test is very 
unlikely under the theoretical distribution, there is reason to reject the established 
null hypothesis* However, the test of significance is not univocal, since the random 
variability of the results of the test can mean that some results, through the random 
effect, deviate significantly from the results anticipated under the established null 

- 88 - 

The principles related to significance tests are illustrated in Figure 1, 

2 1/2 % 

Significant Not significant Significant 

(rejection) (acceptance) (rejection) 

Figure 1 Distribution of the size of the test, if the null hypothesis is genuine. 
The figure also illustrates a bilateral test at the 5 percent level. 

However, on the basis of statistical theory, it is possible to calculate the pro- 
bability that a specific deviation will occur at random. Significance tests are there- 
fore made at different levels of significance, indicating the probability that the de- 
viation could have occured at random. Instead of merely indicating that the results 
deviate significantly from the null hypothesis, it is indicated that the deviations are 
significant at the 5 percent level. 

Finding significance, for example, at the 5 percent level, implies that there will 
continue to be a 5 percent probability that the deviation encountered is the result of 
contingencies. For this reason, if a test of significance at the 5 percent level is used 
to reject the established null hypothesis, there is a 5 percent probability that this 
hypothesis has been erroneously rejected, whereas there is only a 1 percent probability 
that the rejection of the null hypothesis is erroneous if the deviation is significant 
*t the 1 percent level. 

A significance test is often considered as an automatic norm related to the "ac- 
ceptance" or rejection of the null hypothesis. As mentioned above, a significance test 
can never reject with 100 percent certainty a hypothesis, and therefore, the tree breeder 
should take these limitations into account, attempting in each situation to supplement 
the information from the significance test with information on the volume of data material, 
the experience gained from relevant tests, etc. (Brown et al., 1977). 


In the previous paragraph mention was made of the principles governing significance 
tests. We give below a brief description of three tests frequently used, which include a 
fairly large number of the most important methods of analysis mentioned in the present 

- 89 - 

Chi squared test 

Chi squared is shown by the following formula: 

2 (o. - e.) (o ? - e,) 2 <o n - ej 2 
Chi 2 . 1 L + _? 2_ + + __n n_ 

f e l 6 2 

o^ o frequencies observed 

e. e * frequencies expected 

f number of degrees of freedom 

The chi squared test is used to test whether the frequencies observed are in con- 
formity with the established expected frequencies (the null hypothesis). The values, 
calculated in accordance with the above formula, are adjusted to chi-squared distri- 
bution, so they can be tested via theoretical chi squared distribution. 

The so-called "t" test is used to evaluate whether the differences between two 
means can be considered statistically significant at the given level of probability. 

The size of t is calculated according to the following formula: 
fc _ 1 2 where 


n i n 2 

x. and x* m estimates of means 

S * standard deviations 

n, and iu - number of observations forming the basis of x, and x 2 

f number of degrees of freedom (O(n.-l) (n 2 -l)) 

The calculated value is adjusted to the t distribution, which can then be tested 
via the theoretical distribution of Student fl s t. 

F test 

The F test is used to test the homogeneity of two variances. The value of F is 
calculated according to the following formula: 

f l* f 2 * 2 where 

S 2 

2 2 
S. and $2 variances (mean squares) 

2 2 
f-. and f~ number of degrees of freedom for S. and S~ 

The test is based on the so-called F distribution (variance ratio distribution). 

- 90 - 


Before starting the statistical analysis itself, and interpretation of the re- 
sults of tests, certain preliminary investigations should be made, partly to detect 
faults or direct errors in the data material, and partly to verify that the requisites 
necessary for implementation of statistical analysis have been fulfilled. 

Outliers and missing values 

In a first study of the data material it should be checked if there are values 
that deviate considerably from other data (outliers) and that should be assumed liable 
to error. Such investigations can easily be made, for example by tracing each datum 
in a diagram, thus localizing and studying in more detail values which deviate marked- 
ly, if any. If there are large quantities of data and a central data processor is 
available, it might be desirable to use an automatic tracing programme, which is be- 
coming standard equirement in all data processing centres. 

Another possibility is to calculate estimates of the means (for example, means 
of plots) and revise all values which deviate by, say, more than three times the stan- 
dard deviation fron the mean total of the test. Since this involves a fairly large 
quantity of data, the calculations take a long time, and therefore it is easier to 
data process them. 

As well as making the above-mentioned tests of individual values, a check should 
be made to see whether there are any missing values, and generally, whether the num- 
ber of observations per test unit is in conformity with the test plan. 

In a simple design like complete randomization, it does not matter, for variance 
analysis, if there are one or several missing values, apart from the fact that this 
will reduce the number of degrees of freedom for test error. In other designs, it 
may be necessary to replace missing values by estimated values, in order to complete 
the analysis. These estimates, related to different designs, have been demonstrated 
by Freese (1970), Apart from reducing the number of degrees of freedom, missing values 
can be a problem in interpretation of results, especially in fairly complicated designs. 

Tests of requisites 

Most of the usual methods of analysis require that certain requisites be ful- 
filled with regard to data material. Tests of three essential requisites (normality, 
homogeneity of variance and additivity) are described below. 

Normality test 

In view of the fact that many statistical methods are based on the assumption 
that data material is adjusted to normal distribution, it is often useful to have a 
method for the study of this normalcy. 

A chi squared test can be used to test for normalcy, showing whether the distri- 
bution of the data material can be expected to match normal distribution. Snedecor and 
Cochran (1967) give a detailed explanation of the procedure for the test. 

The test mentioned can only shew whether there is conformity with normal distri- 
bution. If not, special tests can be used to determine Skewness and Kurtosis, where 
these exist (Snedecor and Cochran, 1967), 

- 91 - 

Homogeneity of variance tests 

As mentioned above, the F test can be used to test the homogeneity of two vari- 
ances. However, if several variances are involved, Bartlett's test (for the homogeneity 
of a set of variances) can be used. The test has been explained in detail by Snedecor 
and Cochran (1967). 

Variance homogeneity is one requisite for the t test. Is is also a requisite for 
a correct F test in variance analyses. 

Additivity test 

In the statistical models established later in the present article, it is assumed 
that the individual factors of the model are additive, that is to say, that the 
result obtained on one plot or test can be expressed as the sum of the effects of the 
model, for example, effects of treatments or blocks and resudual effects. Turkey's 
test, described by Snedecor and Cochran (1967), can be used to ascertain additivity. 

However, preliminary study of data material will often reveal any lack of addi- 
tivity, since in many cases non-additivity will only be revealed by tracing the data 


If the tests mentioned above show that one or more requisites have' not been fulfilled, , 
it may be necessary to make a transformation of the data material before completing the 
statistical analysis. The following are the three most frequently used forms of- transforma- 

1. Arc sine transformation. This transformation can be used for the values of 
binomial distribution expressed in percentages in the field from 1 to 100. This 
type of data occurs frequently at the nursery stage, when survival, frost damage, 
and other such forecasts are made. 

2. Square root transformation. This transformation can be used in cases where variance 
depends on the mean, for example, in estimating the number of branches by branch 
crowns (Bur ley and Wood, 1976). 

3. Logarithmic transformation. This transformation can be used in situations where the 
variance is proportionate to the mean square. According to Bur ley and Wood (1976), 
as in forecasting flowering rates, using a logarithmic scale (1 1-5 flowers, 

2 6-15 flowers, 3 - 16-35 flowers, etc.). 

A detailed description of the different transformations and their applications is 
given by Snedecor and Cochran (1967). 

Apart from the arc sine transformation of values expressed in percentages, in most 
cases it is not necessary, if the analyses are made on the basis of plot means, to make 
transformations of data. However, the individual researcher should study and evaluate in 
each situation the consequences that the missing requisites, if any, could have for the 
test conclusions. 

Contengency tables 

In many situations, tree breeders face the problem of having to compare proportions 
from two or more independent samples. 

Contingency tables are frequently used, especially in relation to qualitative 
(Mendelian) genetics, to determine whether a specific variation with respect to a quilita- 
tive property is in conformity with what was expected. 

- 92 - 

However, contingency tables can also be of interest in forest tree improvement, 
where quantitative properties are studied in the first place; the tables can be used in 
preliminary investigation of the material, or studies of the different treatments of the 

As an example, a 3 x 2 table is shown below containing results of an investigation 
of two types of planting machines (I and II). The plants are grouped as A * undamaged; 
B - damaged, C - felled. 

A B C T 

Machine I 





Machine II 









On the basis of the selected grouping (A,B,C,) it can be ascertained, through a 
chi squared test, whether a difference is expected between the two types of machine. The 
number expected in each square of the table is calculated beforehand as follows: 

Expected number of undamaged plants (A) for machine I 

155 x 248 

Expected number of damaged plants (B) for machine I - 155 x 61 




Chi squared is calculated as follows: 

(117 - 116. 1) 2 (31 - 28. 6) 2 (15 - 11. 7) 2 , , fl 

116.1 * 28.6 "*" 11.7 " i25. 

It may be seen from a chi squared table with two degrees of freedom ( (3-1) (2-1) -2) 
that the value of 2.38 given there provides no grounds for assuming that the two machines 
in the tests are different. 


Variance analysis is the most widely used method of interpreting the results of 
tests. Variance analysis gives the following information: 

1) estimates of the relative size or significance of each identifiable source of 

2) estimated differences among the materials, the treatments and the sites used 
in the test; 

3) indications of the precision of the estimated differences among the estimated 
means, through their standard error and their limits of reliability; 

4) tests of statistical significance of variances and differences. 

The exact form of the variance analysis will depend both on the test design and on 
the mathematical model used as a basis. Also, both the significance test and the later 
interpretation of the results will depend on whether the individual variables are consi- 
dered ma fixed or random. 

The essential part of all analyses is, however, to estimate variances linked to each 
source of variation (e.g. material, site, residual). 

- 93 - 

Examples of analyses of five different designs are shown below. (SheffS, 1959. 
Searle, 1971). 

Complete randomization 

In this design there is only one source of systematic variation which can be identi- 
fied, i.e., among the treatments. A systematic variation that may exist on the site cannot 
be segregated, unless it is contained in the variation from plot to plot. 

The mathematical model takes the following form: 
*ij ^ * a 4 * EH Where 


individual observation 
p mean total 

1 - effect of the ith treatment 
c lj " residual effect 

A summary of variance analysis table is shown below. 

Source of variation 




Between treatments 

a - 1 

fV'i--*--' 2 


N - a 

1 (x u- x i- )2 


N - 1 

H -.' 

total number of plots in the test 
a number of treatments 
J i number of plots per ith treatment 
Xij individual observation 
*i. estimated mean of ith tretmant 
*' mean total 

The F test appears in the last column of the analysis table. If the F test reveals 
significant differences between the treatments or the materials, the real difference of D 
can be calculated, and the limits of reliability can also be calculated for the individual 
differences among the estimated means. 

The real difference of D (LSD) is calculated in accordance with the following 

VI . S 



' Vff 

t - the value Student's t at a given level of probability, 
S test errors (residual MS (mean square) of the table) 
n number of plots which are the basis of the values compared. 

The limits of reliability for the difference among means may be calculated in 
accordance with the following formula: 

D - ( ai - a 2 ) t + !_ 
V n l n 2 


HI and &2 m estimate of means of a^ and &2 treatments 

t the value of Student's t at a given level of probability 

S - test errors (residual MS of the table) 

n 1 and n 2 - number of observations forming the basis of each of the estimates 
of the means a^ and d2* 

Random block design 

Variance analysis of a random block design is not much more difficult than the 
complete randomization test. There are two systematic sources of variation which can 
be identified; these are blocks and treatments. 

The mathematical model takes the following form: 




individual observation 

- mean total 

effect of the treatment 1 th 

- effect of the block j th 

- residual effect 

(i - 


A summary variance analysis table is shown below. 

Source of variation 






*i < x i.- x ..> 2 




I ^ (X tj - X.J* 



H<vv-v*.> 2 



55 <V X .. )2 

In cases where, at a given level of probability, there are significant differences 
among treatments, the real difference of D can be calculated and limits of reliability 
established for the differences. 

- 95 - 

The test appears in the last column of the variance analysis table. 

Latin square design 

There are three identifiable sources of variation, which are lines, columns and 
treatments, and it should be noted that the number of lines, columns and treatments is 
equal. Through this design it is possible to isolate in the calculations variations 
occuring both between lines and between columns, that is to say, in comparison with the 
random block design, a further division has been made of the systematic site variations. 
As may be seen from the variance analysis table below, the fairly strict design means 
that the number of degrees of freedom for the variance of errors is reduced in compari- 
son with the random block design. 

The mathematical model takes the following form: 


f Ak + fi 

individual observation 
mean total 

effect of the line i th 
effect of the column 
effect of the treatment k th 
residual effect 


a - 1, 

(k = 1, 


A summary of variance analysis table is shown below. 

Source of variation 






X i !..-../ 




I 1 (X - X ) 2 
J J 




I I (X - X ) 2 

K. K. 




9 2 


i 2 -i 

A %K- ...' 

* Since the values represented by i, j and k are aggregated, 
all possible combinations will not be found. 

The F test appears in the table; the real difference of D and the limits of 
reliability are calculated as usual through standard error. 

- 96 - 

Incomplete block design 

Incomplete block designs, particularly lattice designs, are relatively difficult 
to analyze. Since they are used for fairly large and complicated tests, it is best to 
use electronic data processing. 

A brief description is given below of a general "intrablocfc" analysis, which can 
also be used for different lattice designs. 

The mathematical model takes the following form: 

x ij M + a i + PJ + c ij where 

tjj - individual observation 

/Li - mean total 

a - effect of the block i th 

04 - effect of the treatment j th 

fiij m residual effect 

N - I.K. , where K - number of plots per block 

A summary of variance analysis table is shown below. 

Source of variation 






K E (X. - X ) 2 

i JL 

Treatments fcorrected) 


K Q* / X I 








if tx ir x .. )2 

The following comments may be made on the summary: 

The corrected sum (CS) of treatment is calculated in accordance with the 
following formula: 

J * 
A - the number of times that two treatments occur in the same block 

Q J - T J -M n ij B i wh re 

T. - the sum of all the observations included in treatment j 
B. - the sum of all the observations in block i 
n. - 1, if the treatment occurs in block i, a - 0. 

The F test and the calculations of the real difference of D and the limits of 
reliability are made as usual. 

- 97 - 

In lattice designs, Instead of the "intrablock" analysis described above, an 
"interblocks" analysis can be made, also using a variation that can occur between the 
blocks. The fl interblacks ff analysis Is described in detail by Cochran and Cox (1957); 
Burley and Wood (1976) explain the method of analysis in connection with provenance 

The lattice design analysis will be considerably more complicated if there are 
missing values, and in such situations it may be necessary to make the analysis as in 
random block designs, in this case omitting the division into Incomplete blocks. 

Split plot design 

Split plot designs are composed both of main plots and of sub-plots. The main 
plots can be arranged in different designs, but in the example given below it is 
assumed that the main plots are grouped in blocks and arranged in a random block 

The mathematical model takes the following form: 

X-MV" 14 + flj + fl. 4- A-H + t/Jir "*" filk ** -MV where 

A J XV ! i J A, J ^ IV 9 XIV J. J IV 

X^jk individual observation 

jti - mean total 

a - effect of treatment i th (1 1, .. I) 
0.j effect of block j th (J - l f .. J) 
error (A) (k - 1, .. K) 

effect of treatment k th 
lk interaction between A and B 
e ijk * residual effect (B error) 

The variance analysis takes the following form: 

Source of variation 




Treatments (A) 


JK I (X. - X ) 2 

x x 




XK (X , - X ) 2 
J J 

Error (A) 

(i-l) (J-l) 


KFT /Y V .. V X.Y \ 
4*4> IA. .A. A . T X J 

1 J i J i J * 

Treatments (B) 


IJ Z (X - X ) 2 

K. A 


(A) x (B) 

(I-l) (K-l) 

J ?i (x .i fc - x .i.- x ..k +x ... )2 


(Error (B)) 





fff (X iJk- x ... )2 

- 98 - 

As may be seen from the summary, two variances of errors are used to test the 
effects linked to the main plots and those linked to the sub-plots, respectively. 

The real difference of D and the limits of reliability are calculated in the 
same way, using different standard errors for main plots and sub-plots respectively. 

The split plot design is useful in those cases where it is desired to achieve, 
for the effects linked to the sub-plots, greater precision than that desired for those 
linked to the main plots. 


The simple correlation coefiicient is a measure of the degree of linear associa 
tion between two variables. This can often be useful in forest tree improvement, or 
for investigation of the correlation among properties within the same individual, or 
for the investigation of the correlation of a property among different and often 
related individuals. 

The simple correlation coefficient (r) is estimated as follows: 

where SP indicates the sum of the products of the corresponding values of x and y, and 
SC X and SC y indicate the sums of the squares of the deviations for the values of x and 
y respectively. 

The value of r can vary between -1 and +1. The values of -1 and +1 respectively 
indicate complete negative or positive correlations, that is to say, all the points 
included in the bidimensional diagram are found in a straight line with a downwards 
(negative correlation) or upwards (positive correlation) skew. 

The statistical significance of a coefficient of correlation at a given level of 
probability can be determined as follows: 

The null hypothesis: the coefficient of correlation has the hypothetical value 
of Q 0. 

The equation is used to indicate the size of the test. 

If a t table is consulted and the size of the t test with (n-2) degrees of 
freedom turns out to be significant at a given level of probability, the established 
null hypothesis may be rejected. 


The relation between the specific values of a variable and the mean of all the 
Correlative values of another variable that depends on the first, is indicated as a 
regression of the second variable over the first. 

Regressions have the following general form: 

- 99 - 

^i * a + |3x^ + GI where 

y * dependent variable 

a coefficient 

coefficient (coefficient of skew) 

x Independent variable 

e - residual effect 

Coefficients a and f) are estimated as 


a * y -b x 

The significance of linear regression can be tested through variance analysis, 
as shown in the following table. 

Source of variation 






b. SP^ 







1, (y - y.) 2 

In many situations it will be necessary to use a more complicated model, 
including several independent variables (multiple regression) , and in which may 
also be found terms of squares or products or different variables. 


As in variance analysis, there are a series of different models for covariance 
analysis. We will give only one example: a model of complete randomization tests, in 
which, as well as the single identifiable systematic effect, an additional datum has 
been registered in relation to each measurement. 

In covariance analyses an adjustment is made of the values of plots through a 
covariable. In most cases these are adjustments compensating for existing site dif- 
ferences, but they can also be adjustments to compensate for the initial value of 
the material, etc. 

The mathematical model takes the following form: 

1 j 


- 100 - 


individual observation 

mean total 

effect of the treatment i tn (i - 1, .... I) 


co variable 

residual effect 

total number of plots in test 

A summary covariance analysis table is shown below: 

Source of variation 






sc fc - sc ^ - sc 

treat. tot. res. 


x (covariance) 


KcOT "?!^"y yil ~ y "" 2/ 



SC - --, % 2 
res. LL (y - y ) - SC 




sc tot -!?<*- *..> 2 

By including the covariable in the analysis, it is possible, as is shown in the 
table, to calculate a corrected MS error and MS treatment, after which an F test can 
be made in the usual way. 


If in an analysis there are random variables, the estimated MS can be disaggre- 
gated in one or more variability components, each one relating to an identifiable source 
of variation. The following table shows the distribution of expected MS in a random 
block design. 

- 101 - 

Source of variation 


Expected MS (E(MS)) 



*e + J ' a treatment 



'e^ 1 ** block 



a e 



treatment the treatments component 
o block - the block component 
o ^ the residual component 

The Individual variability components can be calculated from the MS values 
estimated In the analysis. 

The most important application of variability components in relation to forest 
tree improvement is the estimation of genetic parameters within populations, also for 
forecasting the genetic gains that could be achieved by selection. 

The present article will not discuss in detail the different genetic parameters, 
since this subject is covered In the article on quantitative genetics. 

- 102 - 


Brown, A.G. & Matheson, A.C., 1977. Statistics: Interpretation of Results, Inter- 
national course in Forest Tree Breeding, Canberra, Australia. 

Bur ley, J. & Wood, P.J., 1976. A manual on Species and Provenance Research with 
particular reference to the Tropics. CFI, Oxford, England. 

Cochran, W.G. fl Cox, M.G., 1957. Experimental Designs, John Wiley & Sons, Inc. London. 

Freese, F., 1970. Metodos Estadisticos Elementales para Technicos Forestales, Centro 
Regional de Ayuda Tecnica, Mexico/Buenos Aires. 

Kempthorne, 0., 1975. The designs and Analysis of Experiments. R.E. Krieger Publishing 
Company, New York. 

Scheffg, H., 1959. The Analysis and Variance, John Wiley & Sons, Inc., New York. 
Searle, S.R. , 1971. Linear Models. John Wiley & Sons, Inc. New York. 

Snedecor, G.W. & Cochran, W.G., 1967. Statistical Methods, The Iowa State University 
Press, Ames, Iowa, USA. 

Putting him on the spot with some difficult questions.... 
(Experimental plantation of Bombacopsis quinata) 

- 103 - 

R.L. Willan 



Objectives * 4 

The need for species and provenance trials 104 

The types, sequence and time scale 104 

Control of trials and their follow-up 105 

Site assessment 105 

Homoclinal comparisons and the choice of species to test 106 

Practical limitations 106 

Phasing and time scales of species trials 107 

Species elimination phase . 107 

Species testing phase 107 

Species proving phase 107 

Phasing and time scales of provenance trials 107 

Range-wide provenance phase 107 

Restricted provenance phase 108 

Provenance proving phase 108 

Plot size, shape and competition 108 

Experimental design ... ... 109 

Cultural treatment and protection 109 

Priorities in field assessment 109 

Recording systems 110 

International cooperation. ... 110 

Summary ... ... ... HI 

References. ... HI 

I/ ^Most of the following lecture has been extracted from "Manual sobre investigaciones 
de especies y procedencias con referencia especial a los tropicos" (Burley and 
Wood 1979). 

- 104 - 


As described in the lecture on tree improvement in relation to national forest policy* 
the objectives of afforestation and hence of species and provenance trials, should be stated 
precisely in advance, in terms of closely defined materials or amenities to be produced. 
The produce required may have a variety of possible uses, e.g. a saw log crop from which thin- 
nings can provide poles and firewood, or it may have a very precise use, e.g. a high quality 
veneer for export. Alternatively, what is desired may be an amenity e.g. decoration, shade, 
shelter or soil improvement. 


Trials are needed whenever adequate information is lacking, either on the requirements 
of the species, or on the characteristics of the site, or both. In such cases embarking on 
afforestation schemes without a carefully planned and executed experimental programme has 
often led to costly failures. 

The choice of species and provenances to use for afforestation involves the extra- 
polation of information from elsewhere. Climatic and ecological matching of a new site and 
the original habitat of a species is rarely enough since it cannot reveal the adaptability 
of the species to new conditions or its ability to grow satisfactorily on a range of sites. 
When information is lacking, the best way to acquire it is through trials of a number of 
species in small plots on representative locations within the area of the proposed afforesta- 
tion project. Provided the locations are carefully selected to sample the range of planting 
sites and are properly looked after, extrapolation of performance from small plots to the 
whole afforestation area should involve far less risk than imprecise comparisons, based on 
inadequate data, between widely separated regions of the world. 

The advisability of species and provenance trials is now generally accepted, but the 
need for their careful planning and for high standards of maintenance and assessment has 
often been less appreciated. The trials themselves can be wasteful and misleading if badly 
planned or executed; and a proliferation of plots, if they are ill-sited, ill-tended and 
ill-protected, is no substitute for a small, wisely planned programme which is tailored 
to the staff and financial resources available. The objective is to derive the greatest 
possible information from a given cost or, put the other way, to obtain the desired informa- 
tion at the lowest possible cost* 

For species with naturally wide geographical or ecological ranges, provenance testing 
is essential. It is easy to be misled in the comparison of species for afforestation if the 
total range of intra-specific variation is not known. 


The ultimate phase is, of course, the complete afforestation project where the source 
populations are reduced to one or two provenances of one or a few species and where the 
annual planting area is reckoned in hundreds or thousands of hectares. It must be recognised 
that there is no standard procedure or time schedule for passage through successive stages 
of testing; nor is there always a need to use every stage. However, the following distinct 
phases are commonly encountered. 

The species elimination phase, the mass screening of a large number of possible 
species in small plots for a short period (1/10-1/5 rotation) to determine survival and 
promise of reasonable growth. The species testing phase is assigned for the critical test- 
ing or comparison of a reduced number of promising species in larger plots for longer 
periods (k-% rotation). The species proving phase is designed to confirm, under normal 
plantation conditions, the superiority of a few probable species. Three similar stages 
apply to provenance testing for species with a wide natural distribution, a range-wide 
provenance sampling phase, a restricted provenance sampling phase and a provenance proving 
phase. Since tnese phases usually apply to species considered likely or promising, plot 
sice and time-scale may exceed those utilised for the comparable phases of species trials. 

- 105 - 


The detailed planning and conduct of species and provenance trials is the responsi- 
bility of the silviculturist or research officer* However, the head of the forest service, 
or the policy-maker to whom the silviculturist is responsible must be capable of evaluating 
the research, and of deciding whether it will answer the right questions, with efficiency, 
and within the resources available. 

To facilitate control the research staff must be required to produce and follow a 
project control plan for each species and provenance trial* This plan must be checked 
and approved before the project is initiated. After any amendments are made the forest 
director should signify approval in writing; this imposes on him the moral obligation 
to do his best to ensure continued finance and staff for the project. If there is no 
promise of continuity in this respect, there is little point in starting the programme* 
Once a project is initiated, however, regular reports should be required (usually annually)* 

A pilot planting project may frequently form an essential intermediate step between 
species/provenance trials and large scale afforestation* It will enable the forester to 
determine optimum cultural and managerial techniques and to make the vital decision on 
whether or not to proceed with the complete afforestation programme. 

It must be remembered that in some cases the results of species and provenance 
trials are applied to the second, rather than the first, rotation* For example, many 
countries are already planting large areas of Pinus caribaea var. hondurensis Barr. & 
Golf., mainly derived from the Mountain Pine Ridge, Belize. Indeed some countries have 
initiated selective breeding programmes based on this material. Nevertheless other 
provenances may be better or contain some valuable genes, and these countries should carry 
out comparative trials to identify such provenances for later planting. There is a danger 
that foresters who have a readily accessible seed source of a "satisfactory 11 species or 
provenance may feel it unnecessary to test others that could be potentially better. 


It is important to realize that each trial does not measure just the performance of 
a given species or provenance, but its performance on a particular site and with a particular 
cultural treatment. Each of these factors interacts with the other two* The recognition 
and mapping of the major site types in a potential afforestation area is therefore an 
essential preliminary to an efficient programme of species and provenance trials* It 
permits the eventual extrapolation of results from the trials to unplanted areas of the 
same site type* 

It is not satisfactory to locate trials by scattering them at random throughout 
potential afforestation areas. Such areas are unlikely to have uniform environmental 
characteristics. In these circumstances the random location of trials may fail to sample 
a widely occurring site type. 

The procedure of assessment should, therefore, be a process of progressive division 
and sub-division to arrive at environmental units useful for the planning and interpretation 
of trials. 

Each unit should be a 'site 1 as defined by Coile (1952): - 

"an area of land with characteristic combination of soil, topographic, climatic, 
and biotic factors". 

The ultimate definitions of these sites will depend on the degree of variability 
encountered. However, a generalized procedure can be followed, as follows: 

Classify each proposed planting area by: 

(i) Latitude - to the nearest degree. (Latitude is related to day length). 

(ii) Rainfall - according to its distribution throughout the year e.g. uniform, one 
dry season, two dry seasons* If the annual rainfall shows wide variation with- 

- 106 - 

in any category a secondary division according to mean annual rainfall will 
be necessary* 

Draw the boundaries of proposed plantations on maps or aerial photomosaics, prefera- 
bly using a scale of 1 : 50,000. On these maps, or a series of transparent overlays, mark 
the following, if known, in this sequence. 

iii) Geological boundaries, or the boundaries of generalized soil groups. 

iv) The boundaries of conspicuous geomorphic features e.g. flood plains, river 
basins, undulating hills, escarpments, dissected plateaus. 

v) The boundaries of major topographic categories e.g. valley bottom, slope, hill 
crest, plateau. 

The ultimate units in the classification are, therefore topographic. 

The close matching of natural habitat and site for species introduction does not 
eliminate the need for trials, since, however accurate the formulae used, the adaptability 
and plasticity of a species cannot be assessed without testing. Moreover, the natural 
distribution of a species may be due as much to the incidence of fire, ecological compet- 
ition or man's activity, as to the measurable features of climate and soil. Many species 
perform strikingly better in a new environment than they do in their natural habitat - 
Pinus radiata and Eucalyptus sallgna are excellent examples of this. 

On the other hand homoclimal information gained from the performance of a species 
as an exotic can be of much greater value, and a review of information from other 
countries can often reduce the number of possible species considerably and lead to the 
inclusion of valuable exotic land races in provenance trials. 

Most obvious non-starters - e.g. Douglas fir at low altitudes at the equator, can 
be eliminated from the start, but in less certain cases it may be better to allow a 
species to eliminate itself in trial, than to eliminate it on theoretical grounds, only 
to have it reconsidered later when these grounds are forgotten. On the other hand, the 
re-testing of species that theoretically should have done well, but did not, is often 
worthwhile, because improved techniques or the build-up of mycorrhizal populations may 
reverse earlier failures. 

It is usually desirable to include a 'standard* well known species of provenance 
in trials, in order to have a reference point against which to judge the performance of 
the unknown populations. 


The size of a programme of species and provenance trials depends upon many factors, 
including staffing, finance and the availability of land, not only for the trials, but also 
for subsequent planting on an operational scale* Security of tenure for the trial areas, 
and public cooperation in the protection of the trials themselves, are essential. The 
availability of suitably qualified staff will govern the kind of work that can be under- 
taken, and training programmes may be a necessary part of the project. The availability 
of transport may be critical in some countries. 

The programme of operations must be worked out carefully in advance and costed as 
far as possible. The total cost must cover not only the initial expenses but also the 
essential maintenance of the plots throughout the duration of the experiments. 

When staff or funds are limiting (and they are almost everywhere) the work needed 
must be limited in advance, and in some circumstances an inadequately replicated series 
of trials may be better than nothing. A precise estimate of species-site interaction, 
on the other hand, costs money and this must be budgeted for. 

- 107 - 

The need to limit the research programme to a practical size makes it essential to 
locate trial plots to ensure: 

(a) Representative coverage of the main site types 

(b) Easy accessibility for maintenance and assessment. 

A larger number of small plots necessitates a long total perimeter, and hence more 
costly protection. A high degree of replication and numerous species involves meticulous 
labelling and supervision which may be difficult with relatively untrained field staff. 
The propaganda value of a set of vigorous plots within sight of a road may be considerable 
and, from the practical point of view, the simpler the statistical design and layout, the 

Species elimination phase 

Object: To compare the performance of a large number of different species on one or a 
number of sites, and to select a smaller number for more intensive trials. 

Features: The individual species unit, or plot, is kept as small as possible. 

The duration of such trials is commonly 0.1 to 0.2 x rotation age, and perhaps 20 
to 40 species could be tested in the initial stages, though continued introductions of small 
number of species are often made over a number of years. 

Species testing phase 

Object: The comparison of a restricted number of promising species, based on previous 
experience, on sites within a broad climatic region. 

Features: Properly designed statistical layouts are particularly important, and plots 
must be of a size to enable reliable assessments to be made up to, at least, 
the first thinning. 

The duration of these trials may be about 0.5 x estimated rotation length. 
Between 5 and 10 species is suggested at this stage. 

Species proving phase 

Object: To confirm, under normal plantation conditions, the results shown by a small 
number of species that have shown themselves superior in earlier phases. 

Features: Plots must be large enough to provide data on growth and yield for the full 

rotation, surrounds must be large enough to eliminate or minimize edge effect. 
In addition to 'normal 1 plantation methods, a range of other management tech- 
niques may need to be tested, always in statistically valid designs. It is 
also appropriate to investigate wood quality at this stage. 


The 'ideal 1 sequence of provenance trials follows very closely that outlined for 
species above. They may be described as the Range-wide provenance phase, the Restricted 
provenance phase and the Provenance proving phase. 

Range-wide provenance phase 

Object: To determine the extent and pattern of variation between provenances (populations) 
of promising species with wide natural variation. 

- 108 - 

Features: Depending on the geographical distribution and variation of the species, 10-30 
provenances are suggested at this stage. It often indicates groups of promising 
provenances, and also areas from which large scale seed imports should be 

This phase is often run concurrently with species elimination or testing. Plot size 
should be small but adequate for a duration of 0.25 to 0.5 x rotation age. 

Restricted provenance phase 

Object: To find sub-regions and ultimately provenances most suited to the sites under 

Features: The differences to be detected between provenances may be relatively slight, and 
experimental design roust take account of this. Generally, 3-5 provenances may 
be expected, with a duration in excess of 0.5 x estimated rotation, using plots 
of the appropriate size. This phase is often run concurrently with species 
testing and species proving phases. Local land races and other derived prove- 
nances should be included where possible. 

Provenance proving phase 

At this stage one or two provenances only will have been selected for each species, 
site and end use. The procedure is the same as that described for species. 


The size of plots, as indicated above, depends on the duration of the trial and the 
expected growth rate of the trees* 

As expense is also an important factor, it should be remembered that trials are 
grown to provide information, and that this may be obtained as well from small plots as 
from large. The single tree plot is of limited use except where a large number of species 
or provenances is being screened for early survival, though it is cheap in space and cost, 
and lends itself to a high degree of replication. Assuming an initial spacing distance 
of 2 to 3 metres the following numbers of trees per plot are recommended* 

Species elimination phase 
Minimum plot size: 5 tree line plot; Maximum: 25 trees (5 x 5). No surrounds. 

Species testing phase 
Plot size 16-25 trees (4 x 4 or 5 x 5) with a 1 or 2 row surround. 

Species proving phase 

Because yield estimates are important, a central plot of 100 trees (10 x 10) plus a 2 row 
surround may be considered a minimum. 

Range wide provenance phase 
25 tree plots, no surrounds. 

Restricted provenance phase 
25-49 tree plots, 1 or 2 row surround. 

Provenance proving phase 
100 tree plots (or more), 1 or 2 row surround. 

Plots are usually square or rectangular, but may need to be elongated to fit 
certain site configurations. 

- 109 - 


This is covered in a separate lecture. The most commonly used design is the Random- 
ized Complete Block. It is simple, flexible and robust, but is less suitable if there are 
very many species or provenances to be compared. 


'De luxe 1 treatment, that is measures to combat local hazards, such as intensive 
weeding or the use of insecticides against leaf cutting ants or termites, should be used 
in order to reduce as far as possible the chance of cultural methods masking the effects 
of other site characteristics. Generally, therefore, competing vegetation should be reduced 
to a minimum and the type of planting stock and method of planting should be the optimum 
for the area concerned. 

Clear prescriptions must be included in the experimental plan on such items as the 
need for fencing or fertilization, time and method of planting, replacement of casualties, 
weeding, protection against fire and pests, pruning and thinning. At planting it is 
essential to ensure that plants are clearly labelled at all times and that planting is 
organized to prevent the confounding of species/provenance plots with individual workers 
whose planting skill may vary. 


Assessment is time-consuming and expensive; therefore the characters to be assessed, 
and the timing of assessments should be laid down in the control plan. 

The initial calibrators of the field stage are assessment of height a few weeks 
after planting (to allow the soil to settle) and a survival count. 

The following are notes on the most important assessments. 


Frequency stage 




Note incidence of pests and diseases 
Identify pests and pathogens. 


1 yr. old. Subsequently 
after climatic extremes, 

100% count. 
Sample count. 

Mean height (h) 

Annual up to about 7 m 
height. Then every 
2-5 years 

Use measuring poles up to 7 m then 
by optical instruments. Accurancy: 
aim at 5%. 100% in initial stages, 
or sample. 

Dominant height (h dom ) 
(mean height of 100 
trees of largest 
diameter per ha) 

Annual up to 7m height. 
Then every 3-5 years 

As above, but less time consuming. 
Accurancy: as above. 

Mean diameter (3) 
at breast height 

1.3 m from ground. 
Annually after crop is 
2-3 m high. 

Diameter tapes most convenient, or 
calipers. 100% sample essential 
Accurancy: aim at 1-2 mm. 

Basal area 

Obtained from measurements of d. 

Stem from Branch site 
and angle 

Start when trees are 7 m 
tall. Then at 3-5 year 

Simple realistic systems best, using 
scoring, 1-7 basis. 

Bark thickness 

Whenever d is measured 
and on thinnings 

Bark gauge, or bark removal on 
thinnings; sample 5-101 of trees. 

- 110 - 


A clear and accurate data recording system is essential for all experimental work. 
Each item of data must be clearly identified and errors in recording and transcription 
occur all too easily. Most commonly, data are recorded by hand on specially designed forms, 
and these should be simple to use and provide a permanent record. The data should be 
analysed as soon as possible, as the results may affect the future management of the trial. 


The transfer of forest reproduction material, throughout the world, has reached un- 
precedented heights during the past few decades; many countries have been trying to inten- 
sify their forest production by introducing new species of provenances (Lacaze, 1978). 

At the same time, there has been an increase in organized research aimed at supplying 
silviculturists with objective information on the vegetative material to be selected. 

An effort on this scale is inconceivable without active international cooperation. 
Some progress has been achieved, but much remains to be done in this field. 

During the past ten years numerous projects have been put into operation, very 
often on the basis of international collaboration. In this connection special mention 
should be made of the work done by various bodies such as the 1UFRO working groups on 
provenance seed collection of North American species, FAO, the Commonwealth Forestry 
Institute, the Danish FAO Forest Tree Seed Centre and the Centre Technique Forest ier 
Tropical (France). 

The number of species on which provenance trials are conducted is very high: just 
as examples, we can list: 

- For species concerning temperature zones: Abies grand! s Picea abies, Picea 
sitchensis, Pinus contorta, Pseudotsuga menziesii, Larix europaea and Populus 

- For species concerning Mediterranean zones: Pinus halepensis, Pinus brut la, 
Eucalyptus camaldulensis, and Eucalyptus dalrympleana. 

- For species concerning subtropical and tropical zones: Cedrela odorata, E. camal- 
dulensis, E. deglupta, E. microtheca, E. tereticornis, E. urophylla, Pinus cari- 
baea, P. kesiya, P. oocarpa, P. pseudostrobus, Tectona grand is, Terminalia ivoren- 
sis and T. superb a". 

In most cases the species studied are exotic to the country in which the experiments 
are being conducted. 

The many experimental plots (past and future) set up using material collected 
within the framework of international efforts are or should be analysed in a cooperative 
manner. This entails an effort of coordination at all stages: 

- Establishment of experimental designs which are, if not identical, at least compa- 
rable with each other in the various zones and countries. 

- Development and utilization of identical procedures for measurement and obser- 
vation. A particular effort should be devoted to the methodology of observations 
on phenologicai stages, pest and disease damage, and measurements of wood quality. 

- Establishment of coordinated timetables for measurements and observations. 

- Utilization of modern means of data processing, at least at regional level. 

In addition to the possibility of checking the results of each experiment against 
those obtained in all the others, this method also has the advantage of making it possible 
to estimate the effects of the interaction between genotype and environment and to iden- 
tify "plastic" population, i.e., those adaptable to a large variety of ecological situa- 

- Ill - 

In addition, such coordinated action also has the great merit of enabling countries 
or regions with limited research, infrastructures or equipment to benefit from the resour- 
ces available and the experience acquired in other countries (Lacaze 1978). 

Recent examples have demonstrated the effectiveness of the assistance that can be 
provided through the implementation of programmes of this kind. 

In countries where there are a considerable number of programmes for species and 
provenance trials and conservation plantations, it is suggested that simple catalogues 
be prepared on the work done and the results achieved, for exchange between specialists* 

Finally, at world level, a publication such as "Forest Genetic Resources Informa- 
tion 11 , edited by FAO, constitutes an interesting initiative for diffusing the most impres- 
sive results of the main programmes. 


The objectives and methodology of trials have been well summarized by Eldridge 
(1979). It is necessary to have: 

(1) Objectives clearly stated; 

(2) Seed collected by a reliable worker, preferably the investigator himself, 
and well documented; 

(3) Seedlings grown in the nursery under uniform conditions and with suitable 

(4) A sound statistical design; 

(5) Plot size determined by the expected variation in the material and by the 
anticipated age of final assessment; 

(6) Planting sites representative of future planting areas and as uniform as 

(7) Great care in labelling and recording at all stages. 

In addition the cooperation of countries in international trials allows a more 
rapid and more efficient accumulation of knowledge at a reduced cost to each country. 


Burley, J. Wood, P.J. Manual sobre investigaciones de especies y procedencias con refe- 
(1979) rencia especial a los tropicos. Trop. For. Paper No. 10, CF1 Oxford. 

Eldridge, K.G. Provenances and provenance trials. In selected reference papers, Inter- 

(1977) national Training Course in Forest Tree Breeding. Aust. Dev. Ass. Ag. 

Lacanze, J.F. Progreso alcanzado en la seleccion de especies y de procedencias. Unasylva 

(1978) Vol. 30 Nos. 119/120, FAO,Rome. 

Coile, T.S. Soil and the growth of forests. Advances in Agronomy Vol. 4: 329-98. 

- 112 - 

Mar cell no Quijada R. 

Institute de Silvicultura 
Universidad de Los Andes 
MSrida, Venezuela 



General remarks 1 ! 2 

Selection of site 11 2 

Assessment of area 113 

Size and density of stand . 113 

Genetic gains 114 

Layout of area 114 

Conformation of stand 114 

Management and upkeep of stand 115 


Seed stands are areas selected in natural formations or plantations to ensure a 
supply of seed of known geographic origin and parentage. 

Seed stands constitute a reliable source of seed of certain genetic quality, 
variable according to the quality of the formation, until such time as seed orchards prove 
necessary and come into production* Since they constitute a local source, usually in or 
near plantation areas in which a certain natural selection has already occurred, they 
offer the advantage of providing seed that is genetically more reliable for the site than 
seed brought from outside or from different environmental conditions. 

Seed stands are a stage prior to the formation of seed orchards. The areas selected 
for seed stands, unlike those chosen for seed orchards, were not originally intended for 
this purpose: they acquire this function owing to certain characteristics of the formation 
and the needs of the plantation programmes. The selection differential is usually lower 
than in seed orchards and the rules concerning the trees present are less strict. 

Genetic improvement through seed stands depends on the quality of the formation 
and on the characteristics under consideration. The trees in the stand are not usually 
subjected to progeny trials, so that their real genetic value is not known. 


The best natural or artificial stands of a given species are located where conditions 
are similar to those of the area in which the seed is to be used. In selecting these 
stands account is taken of the development of the formation (quality of the individuals) 
and of the capacity of the wooded area available to produce enough seed to meet the minimum 
requirements of the plantation programme or programmes. 

In the case of plantations, particularly of exotic species, age is an important 
factor in selection of the site, since the trees must be mature enough to guarantee the 
production of viable seed in acceptable quantities within an acceptable period of tine, 

- 113 - 

but also have deep, wide, green crowns for maximum production of seed in the future. Often 
the areas of physiologically mature or nearly mature trees available are very small; and 
where the areas are sufficiently large, they have been planted only recently. Areas of 
high mortality are usually discarded for seed stand purposes. For some exotic species 
in zones with a difficult environment (such as high altitude or low temperatures, poor 
soils or low rainfall), there is the problem that even when the species grows, repro- 
duction is seriously affected, which is one constraint to the establishment of seed stands* 


Once the area has been selected, the phenotypic quality of the individuals and the 
initial average production are assessed morphologically and quantitatively. This will 
enable us, in addition to laying down criteria for establishing the seed stand, to have a 
basis for calculating the genetic advance to be expected in plantations grown from the 
seed produced in that stand. 

Morphological assessment is based on a few characteristics -3 to 5- considered to 
be of importance for use of the plantation product. For timber and wood-based products 
straightness of the trunk, forking and anomalies (particularly in conifers) are frequently 
used, with different scales or categories. Some characteristics may be decisive with only 
two criteria: rejection or acceptance. For example, trees with visible signs of attack 
by certain insects or diseases are completely rejected, because of the ease with which 
susceptibility can be inherited and the decisive importance of this character for the 
success of future plantations. For the other characteristics, it will depend on the 
quality of the formation how rigorous the selection is. 

Quantitative assessment is made by measuring height and diameter and evaluating 
the form and quality of the trunk. 


For a given quantity of seed required in a given programme, the size of the area 
will depend on the productivity of the species and the density of the stand used. 

The productivity of a species is given by the number of viable seeds per tree. 
This is dependent upon the number of fruits per tree, the number of seeds per fruit and 
the viability of the seeds. Prior sampling is essential to provide an estimate of 

The density of a stand depends on its age and/or development. In plantations, 
between 100 and 300 trees per hectare are generally used, depending on the species and 
on the environmental conditions. Competition between the crowns is always avoided in 
order to ensure the maximum number of reproductive buds. 

In tropical natural areas the possibilities of establishing seed stands are limi- 
ted owing to the heterogeneity of the stands and the limited number of each species per 
hectare, except in some specific types of forests. 

Seed stands are characteristically monospecif ic, since the movement of pollen 
between trees of a species is facilitated if the other species are eliminated. However, 
in natural stands the possibility could be considered of combining two or three taxo- 
nomically very different species so as to avoid problems connected with hybridization. 
In this way more effective use can be made of an area devoted to producing seed. 

For reasons of cost and productivity it is considered that an area of less than 
5 hectares is not suitable for the commercial production of seed. Above this minimum, 
the size will depend on the requirements of the programme, the original area covered by 
the species, and the resources available for effective management of the area. 

- 114 - 


The main value of a seed stand is that it constitutes a local source of seed of 
sufficient quality and quantity to minimize or eliminate dependence on external sources 
that are quantitatively and qualitatively less reliable. 

Any gain that may be achieved will depend on the quality and original conforma- 
tion of the stand. The more regular the distribution of qualities among individuals in 
the stand, the more opportunities there are for a selection that will lead to an impor- 
tant genetic improvement, since there is more likelihood of all the desirable combinations 
being achieved in one or few individuals. The more uniformly one or another type of 
quality is distributed, the less advances are to be expected. 

In first-generation plantations, particularly of exotic species and under some- 
what limiting environmental conditions, distribution is often observed to be slanted 
towards poor quality, with very few good individuals. In this case the greatest gain 
to be hoped for is in survival, with a slight or regular improvement in development 
quality. The improvement in survival is explained by the fact that the seed comes from 
individuals that have survived well for a first-generation plantation, showing a certain 
degree of cultural adaptability to the site, demonstrated by the fact that they have 
achieved an acceptable stage of development under certain limitations. 

Improvement in development is expected because, although the structure of the 
stand and the need for a minimum number of trees per hectare may lead to the selection 
of trees that fall outside the category of "good", the latter will be relatively more 
frequent in the new (select) population and hence have a greater genetic impact. 

A seed stand consists of a production area, or effective area, and a buffer area. 

The production area is the area where the seed will be collected and which may 
receive the cultural operations needed to promote production. 

buffer area serves to limit contamination by pollen from uncontrolled external 
sources, since seed stands are usually established in plantation zones. The buffer is 
composed of the same species as the stand and therefore contributes pollen to the pro- 
duction area. The same criterion regarding the effects of selection of individuals is 
used as in the production area, but no cultural operations are undertaken. 

buffer area will be between 100 and 500 m wide, depending mainly on environ- 
mental factors that favour dispersal of the pollen. 


Once the area has been selected and assessed, the guidelines for intervention 
and conformation of the stand are established. Individuals should as far as possible be 
distributed regularly over the whole area, with a minimum space between them to encourage 
flowering and pollen dispersal. The minimum space is that necessary to avoid precocious 
competition and may be as much as 1.5 times the initial spacing in young stands. In 
any case, spacing will depend on conformation of the neighbouring trees and the density 

In plantations work can be done systematically by rows, thus encouraging regulari- 
ty of the resulting stand. Since the aim of selective thinning is to leave only one quar 
ter to one sixth of the original crop per hectare, it is often done in stages (2 to 4 
years), so that any necessary corrections can be made to the initial operation. 

- 115 - 

of the area the first year and the remaining three quarters in the other two years. 

The time taken to establish a stand usually depends on the area to be covered and 
the constitution of the initial formation (density and size of trees). 

It is essential that the project leader mark the trees to be operated on so that 
at the same time as he/she trains the auxiliary staff who usually complete the work, the 
necessary correction can be made. Technical supervision is necessary in any case in order 
to guarantee good results. Marking should always be done well in advance of tree clearing, 
and this should be taken into account in planning activities. 

Felled trees must be removed from the stand in order to eliminate obstacles and 
possible sources of attacks by pathogens. 

The stumps should usually be fumigated to prevent them becoming foci for insects 
and diseases. 


The main management activity in a seed stand is thinning. Protection against fire, 
insects and diseases is also of great importance. The maintenance of firebreaks and 
adequate access roads permits effective mobilization in the event of fire. The elimination 
of low vegetation and control of stumps also helps to avoid possible centres of infection. 

Biological and genetic research work, such as studies on the use of fertilizers 
or other means of encouraging flowering and fruiting, can also be conducted in seed stands. 

- 116 - 


W.H. Barrett 
Fiplasto S.A., Buonos Aires, ARGENTINA 



Introduction 1 16 

Background and present situation in Latin America , 117 

Location and selection of seed stands 118 

Selection criteria 118 

Characteristics of the stand 119 

Age 119 

Area 119 

Isolation 119 

Accessibility " ... 119 

Stand management 119 

Seed production 120 

Genetic gains 120 

Bibliography 120 


The use of seed stands is an interim measure in most forest improvement programmes, 
which^enables improved seed to be obtained in a short period of time. It is considered an 
interim measure because the intensity of selection from stands can never be the same as 
that from seed orchards. There will therefore be less improvement, or rather, fewer genetic 
gains . 

In regions where large-scale afforestation programmes are undertaken, as is usual 
in tropical or sub-tropical areas, there is a great demand for seeds. The lack of sufficient 
quantities of seeds is often a bottleneck in plantation planning. 

The source from which the seed is extracted, (its provenance) is important to the 
future of the plantation. This source will influence not only the volume to be obtained, 
but also adaptation to the plantation site, plant health, type and rate of growth, wood 
quality, etc. 

Seed stands can be selected both in native forests within the range of natural 
dispersion of the species, and in plantations within and outside the area of origin. 

This paper is based on the lecture given by H. Keiding at the FAO/DANIDA Training Course 
on Forest Seed Collection and Handling, Thailand, 1975, with additional information from 
H. Barner, from the FAO/DANIDA Training Course on Forest Tree Improvement, Kenya 1973. 

- 117 - 

Various names have been given to seed stands, particularly in English-speaking 
countries, such as "seed source", "selected stands", "seed stands", "seed production areas" 
which are discussed in detail by Barner (1973). A distinction must be made between stands 
selected for the general quality of the trees and stands in which intensive thinning has 
been carried out, eliminating inferior individuals. 

The following definitions are adopted in accordance with the OECD (Organisation for 
Economic Cooperation and Development) scheme: 

Stand; "A population of trees possessing sufficient uniformity in composition, 
constitution and arrangement to be distinguished from adjacent population". 

Seed stand/Seed Production; "A plus stand that is generally upgraded and opened by 
removal of undesirable trees and then cultured for early and abundant seed production". 

The objectives of creating seed production areas of seed stands are, according to 
Matthews (1964): 

1) To proudce seed of improved quality through selection and elimination of inferior 
trees, favouring those that are vigorous, straight-stemmed, healthy and producing 
wood of good quality. 

2) To concentrate seed collection in a few specially treated parts of the forest, 
thus making seed collection easier to organize and control. 

3) To improve the germinative energy and germinative capacity of the seed collected. 


The selection and establishment of seed stands is the quickest method of obtaining 
any improvement in the quantity and quality of seed harvested. It would appear that integral 
and long-term improvement procedures include the selection of seed stands to obtain improved 
seeds, until seeds can be obtained from more carefully selected trees in seed orchards. 

At present, seed stands are sometimes still the only way for some countries to obtain 
seed for their plantation programmes. This was the practice most frequently used in the past. 

Stands are usually selected from mature forests, and subsequently treated to trans- 
form them into seed producing areas. 

Forest gains or improvements, and the efforts needed to achieve them, will depend 
on the nature of the basic material one has to work with. 

Following Bamer (1973) and Jones and Hurley (1973), the basic u>aterial has been 
classified in accordance with its origin and location, as follows: 

1) Mainly located within the range of the species, whether in natural stands or 

This group, particularly the conifers, includes such countries or regions as the 
United States, Canada, Mexico and countries of Central and Southern Europe* 

2) Mainly exotic species, introduced into the region, where they have been cultivated 
fairly extensively. 

This group can be found in almost all the countries whose forest production is based 
en extensive cultivation of exotic species. 

Conifers, particularly, nan be found in countries such as New Zealand, Australia, 
Chile and Spain, where Pinus radiata has been extensively cultivated; in South Africa, 
Zimbabwe, Malawi and Kenya (Pinus patula); in Argentina, Brazil and Australia (pines froip 
the southeastern United States); Great Britain and central western Europe (Pseud ot suga menzlesii 
and other species of pines), to give only a few examples. 

- 118 - 

3) Mainly in experimental plots of uncommon or recently introduced species. 

This group includes most tropical and sub-tropical countries, and also many species 
now being tested in developing regions. It applies particularly to species trials and 
prevenances within species which have performed outstandingly in trials as compared with 
the material widely used throughout the region. 

As regards the use of seed stands of native species in tropical regions, the limited 
use of this system can be explained by the fact that native species are usually replaced by 
fast-growing species which are almost always exotic. Even when some native species can be 
used, the mixed nature of the forest makes it very difficult or practically impossible to 
transform a forest into a seed stand. 

Tropical areas in Latin America are no exception to this rule for native species. 
Species are scattered throughout mixed forests, where a large number of species, mainly 
hardwoods, grow side by side. Native conifers occur in climates ranging from sub-tropical 
to cold temperate. Apart from some exceptions, such as certain pine species in Mexico and 
Araucarias in Argentina, Brazil and Chile, most conifer species are not cultivated and are 
Araucaria angustifolia. In the Argentine province of mi si ones alone, scape 30 000 hectares 
have been planted. So far, seed has been collected from native stands - although seed 
stands are now being developed. 

Group 2, as mentioned above, includes Chile, Brazil and Argentina. The questionnaire 
to be completed by course participants will probably provide further inf anna t ion of species 
cultivated in other American countries. 

The remaining species used on a large scale in many American countries are still 
included in group 3, since there are no forest plantations old enough for selection of seed 
stands. The oldest and most successful experimental plantations are used for this purpose. 


Here again, a distinction must be made between native species in their native area 
and exotic species widely cultivated and recently introduced. 

In the first case, selection will have to be based fundamentally on existing infor- 
mation from the various provenance trials which determine suitability for climates and 
soils. Chce the region of area has been determined, a stand will be selected for optimum 
wood quality and quantity, and quality and regularity in seed production. 

As regards exotic species, a distinction must be made between those widely cultivated, 
such as Pinus radiata, . caribaea, . elliottii and IP. patula in the southern hemisphere, 
and those not widely cultivated or only recently introduced. The methodology indicated 
below refers to the first group, where selection of stands will have to be stretched as far 
as possible using all available information on the value of the seed source, reproductive 
quality of the stand under consideration, results obtained in previous plantations, etc. As 
regards the second case (species not widely cultivated), progress will have to take its own 
course, little can be done about it. 

The total area needed for seed stands for any given region depends on the annual 
demand for seed, productivity of stands, by tree or by area, and quality of production. 

Selection criteria 

In selecting the seed stand, its uniformity and volume production must be borne in 
mind, and the trees must have well-shaped trunks, good growth habit, quality of wood and 
good health. 

In European countries (for example, Sweden) stands are graded into "plus", "normal" 
and "minus", there being a tendency to subdivide the plus stands into different categories 
depending cm their degree of superiority. In other countries (for example, Great Britain), 
the stands is classified on the basis of individual trees. Samples are taken systematically, 
within which all the trees are measured and classified. The total of points in the simple 
determines the total points of the stand, according to which it is declared plus, almost 
plus, normal or minus. 

- 119 - 

Although they appear different, selection criteria in both cases are similar with 
respect to the characteristics chosen to classify the trees. Selection criteria are also 
similar to those of individual phenotypic selection, but they differ fundamentally in the 
intensity of selection. 

Characteristics of the stand 

Age varies with species and region. A stand should be old enough to have proved the 
value of its seed in areas where it was used before. It should be mature enough for good 
flowering and fruiting. It should also be old enough to permit correct evaluation of stand 
characteristics. Barner (1973), however, maintains that, due to possible difficulties in 
evaluating some characteristics in older stands, it is desirable to start evaluation at an 
early age, if the stand promises to be valuable, and keep this information on file and up 
to date until needed. The seed stand, although it should be mature, should not be very old, 
so that seed extraction can continue for a reasonable number of years. 


It is not possible to establish the area of a seed stand in advance. In general the 
criteria in the minimum area in which it is economically worthwhile to collect seed. In 
western Europe, 2 to 5 hectares are considered the minimum area for seed stands. The stand 
should also contain enough individual trees of the desired species to allow for reasonable 
cross pollination. This is particularly important in mixed forests. 


The isolation factor should be borne in mind, to ayoid contamination by inferior 
stands. Wind -pollinated conifer species are more difficult to isolate than broad leaved 
species, which are usually insect-pollinated. Isolation can be achieved by choosing stands 
which are between 300 and 1 000 metres away from any possible source of contamination. In 
some cases, plus stands can be found sufficiently far away from contaminating minus stands 
for the forest itself to be used as the buffer. In this case, only the area of the plus 
stand farthest from the contaminating source is harvested, using part pf the same stand as 
a buffer zone. 


Normally, easily accessible stands should be selected, to reduce the cost of manage- 
ment, inspection, upkeep and seed collection. When genetically adequate species are found 
off. the beaten track, it is preferable to collect their seed and sow it in places where 
it is easier to reach. 


The treatments normally given to a stand within an area to be used as a seed source, 
and to convert it into a seed stand, are as follows: 

1) Removal of minus trees to improve the genetic quality of the seed. 

2) Thinning, to allow adequate space for flowering, fruiting and seed collection. 

3) Clearance of the undergrowth to facilitate seed inspection and collection. 

4) Demarcation of the stand, particularly when there are contamination problems. 

5) Treatments to increase production, such as pruning or application of fertilizers. 

6) Other treatments to protect fruiting, such as application of fungicides and 

To improve management of seed stands, it is desirable to keep a record covering all 
activities and treatments applied, pheno logical information, and information on 

- 120 - 


Relatively little is known about seed production from a seed stand , as compared with 
a stand that has not been treated. It is known that when a stand is thinned and the canopy 
opened, an increase in seed production is immediately obtained. However, the duration of 
this effect is not known. It is probable that there is an increase in production per tree, 
but it is not known if there is greater production per area. An improvement in seed quality 
is certainly achieved by eliminating minus trees and thinning to get more light to the 
crowns of selected trees. 

Nor is there any concrete information on increased production as a result of 
fertilization, since almost all research has been done on seed orchards. In Japan, studies 
made en Pinus densiflora and Larix leptolepis showed that the application of fertilizers 
improved seed quality, and thinning produced an increase n seed quantity. The best result 
was dbtained with light thinning; this was not only better than the control treatment (with- 
out thinning), but also better than heavy thinning (Asakawa et al, 1969). 


To date, there has been an increase in the quantity and physiological quality of the 
seed produced in seed stands. 

In species with wide variations in breed or ecotype, usually species with a large 
range, maximum gent ic gains of the seed stand depend on the correct choice of the basic 
population. The original source of seed should be known through previous studies of prove- 
nance trials or as a result of intensive and extensive silviculture. For this, both the area 
of origin and plantations established in the region with material from the same source can 
be used. For example, in Argentina in provenance trials it was found that populations of 
Pinus taeda from central Florida with 7-year rotation grew 200% in volume as compared with 
populations from the foothills of Georgia. Something similar happened with the same species 
in Queensland. As the basic population used for seed stands is from different sites in 
Georgia, the production in volume can easily be improved by using seeds from Florida stands. 

As regards characters within the population such as shape, width of crown, density 
of the wood, etc., this will depend directly on the her it ability of the character and the 
intensity of selection. This being a mass selection, with relatively low selection intensity, 
genetic gains will depend on heritability. 

Okie example from Shelbourne (1969) for diameter low (heritability) and bole straight- 
ness (high heritability) in Pinus Radiata at a selection intensity of 1 in 10, expressed in 
percentage of the average population before thinning, is 25% for bole straightness and 5.6% 
for diameter. At a selection intensity of 1 in 20, the gains anticipated are 29.2% and 
6.7% respectively. 

In Latin American countries, where many factors relating to species, provenance, site 
and their interactions, have not yet been explained, much greater gains may be hoped for in 
the search for species and provenance better suited to each region. For this reason a tech- 
nique recomnended is that of establishing simultaneously with species and provenance trials, 
large-scale plantations of each of the biological units tested. This material could be used 
in the future as a source of seeds in the short term, and selections could be made from it 
for long-term improvement programme. 


Asakawa, S. & Keiji Fujita, 1969. Studies on the management of experimental seed stands pf 
Pinus densiflora and Larix lep to lapis and the results obtained for three years (1962- 
1964). Bull. Gov. For. Exp. Sta. No. 184, Tokyo, Jap. 

ftarner, H. 1973. Classification of sources for procurement of forest reproductive material. 
FAO/DANIDA Training Course on Forest Tree Improvement, Kenya, pp. 119-121. 

E lias on, E.J. 1969. Development of seed production areas-an interim selection* FAOHEUFRO. 
Washington. 2: 1367-1371. 

- 121 - 

Jones, N, & J. Bur ley, 1973. Seed certification provenance nomenclature and genetfc history 
in forestry. Silvae Genetica 22 (3); 53-58. 

Keiding, H., 1975. Seed stands. FAO/DANIDA Training course on forest seed collection and 
handling - 2: 192-211 - Thailand. 

Matthews, J.D., 1964. Seed production and seed certification. Unasylva 18 (2-3): 104-108. 

Mittak, W.L. , 1979. Cursillo para el mane jo de rodales seleccionados para la produccipn de 
semi lias. Institute Nacional Forestal BANSEFCR Proyecto GUA/78/005. Guateipala. 

Shelbourne, C.J.A., 1969. Predicted genetic improvement from different breeding methods. 
Second World Consultation on Genetic Improvement. FAQ-IU7RO Washington 2: 1023-1029. 

- 122 - 


C. Palmberg 

Forest Resources Division 
Forestry Department, FAO 



Introduction 122 

Stand selection 122 

Age 123 

Area 123 

Isolation 123 

Stand Management * 123 

Bibliography 123 


The purpose of a seed stand is to meet immediate needs of relatively large geneti- 
cally somewhat improved quantities of seed by setting aside stands of not lower than 
average quality and managing them with the primary objective of seed production. 

General principles of selection and management of seed stands, as outlined by 
Barrett in the previous lecture, are valid for hardwoods as well as for conifers. Points 
on which methodology between hardwoods and conifers differ or which may require particular 
attention when dealing with hardwoods, are outlined below. References to publications 
dealing with the establishment and management of hardwood seed stands are listed at the 
end of this note. 


Hardwoods are often dioecious. In selection of stands, the availability of both male 
and female trees must be ascertained and when thinning, care must be taken to leave 
a balanced sample of male and female trees; 

In species with pronounced root-succering ability large areas of trees which are 
genetically identical sometimes occur. For these species the seed stand approach may 
not be suitable; 

When dealing with plantation species it is very important that origin of the stand is 
known. Exotic plantations may be based on seed from relatively few trees and although 
the stand itself may be of superior growth and form, the genetic quality of the seed 
collected from it is likely to be poor due to inbreeding effects. Sometimes, e.g. in 
the case of eucalypts, the stand may be of hybrid origin, and seed collected from F, 
hybrids and from subsequent generations will produce unacceptably variable stands 
because of segregation of the genes; 

- Selection criteria such as health of trees, yield and form are universal. However, 
additional, species-specific selection criteria will often be included for hardwood 
species; examples of such criteria are selection against genetically controlled 
defects characteristic of some species, such as the tendency to produce epicormic 
shoots having bud traces extending to the pith as in liquidarabar styracif lua and 
against the formation of a large number of kino veins in Eucalyptus regnans; se lect ion 
for extremely low wood density in Ochroma lag opus selection for specific wood 

characteristics such as attractive colouring or patterns of wood in some high- 
quality cabinet timbers; selection for high coppicing ability especially in some 
species grown on short rotations. 

- 123 - 


The stands must be young enough to be able to respond to the first thinning by the 
development of crowns capable of producing large seed crops, but old enough to give some 
indications of the characteristics to be selected for. 


The distances between individual trees of the same species are often large in forests 
with a heterogeneous species composition. Provided these species are outbreeding and the 
seed stand appearance therefore per s.e is valid, the areas needed are much larger than if 
dealing with forests consisting of one or only a few species. If the density of the species 
is very low, the seed stand approach cannot be used. 


Knowledge of the breeding system of a species is fundamental. If a species is in- 
breeding, no isolation is naturally needed. However, forest tree species are largely out- 
breeding, possessing various mechanisms to prevent selfing. 

Effective pollen flight distances and the distance pollen will be transported by 
insects are not well documented for many hardwood species. The isolation of hardwood seed 
stands should be kept similar to that for conifers (- 300 m) until further studies have been 


Seed stands should undergo a series of thinnings to increase quantity and improve 
quality of the seed produced. The first thinning should be made before serious crown 
competition sets in; the main purpose of this thinning is to allow the development of wide, 
deep and dense crowns, i.e. to increase potential quantity of seed at the same time removing 
inferior phenotypes. During the development of the stand, further variation will manifest 
itself; as phenotypic selection will generally be more reliable in mature trees, the im- 
provement achieved through thinnings will gradually shift its main emphasis from quantity 
to quality. 

Knowledge of the biology of flowering and seed production is essential for the mana- 
gement of seed stands, and this knowledge is often scarce or lacking in hardwoods. It is 
important to be able to forecast the exact time of the year that flowering will take place 
so that thinnings can be timed to precede flowering any one year; incorrect timing may 
cause a delay in the benefits of the thinning equal to the period between successive seed 
crops. It is also essential to know the time required between pollination and the matura- 
tion of seed so that the relative degree of improvement, which is directly related to selec- 
tive thinnings, can be forecast. 


Melchior, G.H., 1977. Programa preliminar de un esayo de precedencia de Cordia alliodora, 
Cupressus lusitanica y otras especies nativas y ex6ticas. Proyecto Investigaciones y 
Desarrollo Forestales COL/74/005. PIF no. 7. Bogota, Colombia. 

Melchior, G.H. & Venegas Tovar, L., 1978. Propuesta para asegurar el suministro de semillas 
de Eucalyptus globulus en calidad comercial y geneticamente majoradas. Proyecto Investiga- 
ciones y Desarrollo Forestales COL/74/005. PIF no. 14. Bogota, Colombia*. 

van Dijk, K. , Venegas Tovar, L. & Melchior, G.H., 1978. El suministro de semillas como 
base de reforestaciones en Colombia. Proyecto Investigaciones y Desarrollo Forestales 
COL/74/005. PIF no. 13, Bogota, Colombia. 

- 124 - 

N. Quijada R. 

Instituto de Silviculture 

Universidad de Los Andes 
Mfirida, Venezuela 



Genotype and environment 124 

Factors to be considered in the selection of individual trees 125 

Type and number of characters 125 

Intensity of selection 125 

Propagation of plus trees 126 

Selection procedure 126 

Method of improvement through selection . 126 

Sources of trees for selection 126 

Selection criteria 127 

Selection terminology . 127 

Bibliography 127 

Annex 1; Example of selection criteria: selection criteria for seed stands ... 128 

Annex 2; Example of selection criteria: selection criteria for seed orchards .. 129 


In practice, the commonest improvement methods are based on the selection of indi- 
viduals meeting certain minimum requirements for specific characteristics which are impor- 
tant for some specific purpose. 

The external appearance, or phenotype, of an individual is the first guide for the 
breeder. Although the phenotype is based on 2 components: genotype and environment, which 
may or may not be of equal importance in the resulting appearance, the breeder bases him/ 
herself initially on the probability that a good phenotype has a sufficient genetic basis 
to produce a reasonably favourable reaction in different environments. 

Many characteristics are known to be genetically controlled, either partly or almost 
entirely, so that external appearance reflects inherent potential to a large extent. Other 
characters vary more in this reaction to environmental conditions, so that a specific 
phenotype reflects only one of the many possible reactions. 

The existence of different forms or alleles of a gene, and the occurrence of poly- 
genesis, pleiotropy epistasis, etc. result in great complexity in the expression of a 
character or various combinations of characters. The breeder selects from the whole range 
of expressions, or variations, the most favourable types or combinations. The selection 
practised by the breeder is controlled or truncated, in that he/she selects towards an 
extreme of the original population and tries to shift the mean of the future populations 
in that direction. 

-125 - 

The importance of the genetic component of a character can only be determined by 
field trials, either through vegetative propagation of the individuals selected in 
different environments, or through their progeny (progeny trials). Meanwhile, external 
appearance will continue to be the basis for work. 

The stricter the selection criteria, the more guarantee there will be of good genetic 
gain. However, this depends on the species, the product and the quality of the population 
available for selection. 

An initial survey of the areas of distribution (plantations or natural forests) helps 
to establish a criterion for the requirements of the various characteristics. This is 
particularly important for selection programmmes with first-generation exotic or native 
species affected by dysgenic selection over much of their range. 


The success of selection based on phenotype expressed as genetic advance or gain, 
depends on various factors, in particular the type and number of characters in the 
selection, the intensity of selection and the method of propagation. 

Type and number of characters 

The type of character (high or low heritability) has a decisive influence on the 
advance that can be achieved. Highly inheritable characters are easier to manipulate and 
more predictable in their responses. Such characters are straightness of the trunk, 
forking, and resistance to diseases. Characters with low heritability are less predictable, 
since they require greater environmental control. They include important economic character- 
istics such as physical and chemical properties of the wood. 

The number of characters also affects the response obtained. It has been shown that 
the greater the number of characters involved the more difficult it is to obtain advances 
of specific individual characters. This is due to two factors: first, different characters 
have different inheritance patterns that require different selection intensities. If the 
number of individuals possessing a given character is increased, they will be increased 
danger of introducing undesirable phenotypes of this character. Secondly, different 
characters may be inversely correlated, so that greater strictness in one would have a 
negative effect on the expression of the other. 

It is therefore necessary to concentrate on a few characters at a time. First, one 
chooses the most manipulable and the most important selection factors, such as straight 
bole, forking, vigour, etc., and second, one considers the properties of the wood (5). 

Intensity of selection 

Trees are selected with two considerations in mind, among others: that the trees 
selected have as little family affinity as possible, in order to avoid problems connected 
with consanguinity; and that it be possible to obtain a minimum number of individuals, in 
accordance with the purpose of the selection. As regards the first, in natural stands 
distance is a reliable indicator. The closer the trees, the greater the probability of 
consanguinity, simply because of how pollen moves. Widely spaced trees, either in the 
same area or, better still, in different areas, show fewer family affinities. In plan- 
tations we expect at least the trees planted in the same year to be more related to each 
other. From one year to another, and one site to another, the specific source of seed 
may vary, particularly when very large lots of seed are required. In this instance an 
effort will be made to reduce the total number of selections on any one area of the 

On the other hand, the number of selections influences the range or genetic base. 
A limited number of selections will create a very narrow base, which could rapidly lead to 
problems, including problems of consanguinity. In seed orchards 20 trees are considered 
an absolute Minimum for maintaining a sufficiently broad genetic basis to achieve important 
advances, and to ensure that the seed produced can adapt to the natural variability of 
plantation sites. This is based on the assumption of regularity in time and quantity of 

- 126 - 

flowering and fruiting of all trees. Since this does not always happen it is preferable to 
use a greater number. In seed areas the number of trees left will depend on seed require- 
ments, the development of the plants and the area available, but it is usually between 150 
and 250 per hectare. 

Strictness with regard to intensity of selection diminishes as parameters of varia- 
bility, such as heritability, become available. Intensities of 1 tree per 700 hectares and 
1 tree per 1.2 hectares have been used for Pinus radiata in New Zealand and Australia 
respectively; 1 tree per 8 000 for Cupressus lusitanica in Colombia and 1 per 750 (1 per 
0.65 hectare) for Pinus caribaea in Cachipo, Venezuela (4, 5, 8). 

In any case the final decision will depend on the variability of the species and on 
the immediate needs for seed in terms of quantity and quality. 

The intensity of selection is determined in various ways. One is by means of a 
selection differential expressing the difference between the mean of the original popula- 
tion and the mean of the selected trees. 

The mean of the original population can be estimated by sampling the population. 
Since this may be expensive and time-consuming, the means of the 4 or 5 best trees near 
the plus tree have been used as a population reference model. 

With this value, and with the estimated heritability value (h ) for a particular 
character, we can calculate the genetic gain (R) expected. (See the paper on quantitative 
genetics for more detailed information). 

Propagation of plus trees 

In the case of trees selected for inclusion in seed orchards, there are two 
alternative methods of propagation: sexual and asexual: which determine the type or orchard 
to be formed (seedling and clonal respectively). 

Method of improvement through selection 

The two most common methods of selection are stand selection and selection by 
family (1, 10). 

In stand selection, individuals are selected on the basis of phenotype and then 
allowed to cross-fertilize freely. This is the usual practice in seed stands and seed 
orchards, where the seed is mixed without concern for family relations. To test the 
effectiveness of stand selection, a progeny trial may be conducted with groups of both 
plus and normal trees from the original populations. This procedure has proven effective 
for highly inheritable characters. 

Selection by family makes it possible to control parental relationships in the 
resulting progeny, thus facilitating continuous evaluation of the plus trees. The most 
common family relationship is sibs or half sibs. Different procedures and cycles of 
assessment, grouped under the generic term of recurrent selection, permit the elimination 
of trees originally selected and the incorporation of new selections. The practice is 
most common in seed orchards for production and control of development (with progeny 

Sources of trees for selection 

The source of the trees to be selected is determined by the use to be made of the 
selections. In the case of seed stands, selection is made within a specific area, normally 
a plantation of known origin, but sometimes a natural stand which has good overall 
characteristics of vegetative and reproductive development and is large enough to guarantee 
sufficient seed output. It must also be located in the zone where the seed will later be 
used. For seed orchards, the total area of natural and artificial stands within a climatic 
region is taken as the reference. In order to have a good initial basis for work, initial 
election can be in the best stands of the best provenances. 

- 127 - 

Selection criteria 

There are two practical methods for assessing trees: individual evaluation and 
comparative evaluation. The first method consists of evaluating each tree on its own 
merits, according to scales of values for the classes in the characteristics. The different 
classes in each individual characteristic will be given by technical criteria, in accordance 
with the phenotypic variants discernible in each one. This alone establishes a certain 
level of sublectivity. This method of evaluation is appropriate in samples made in for- 
mations that are to be used for seed stands, and in which it is desired to assess the 
formation in order to intake decisions on the intensity of and criteria for interventions. 
It is also used to evaluate the development of plants in progeny trials. In this system a 
level is established beneath which any tree is automatically eliminated, regardless of the 
values in other characteristics. 

The second method uses scales of values resulting from the superiority of the 
candidate tree over comparable trees in the vicinity. Usually a number of additional points 
are assigned for so much percent or absolute value of superiority, the magnitudes depending 
on the weight it is desired to give to each characteristic. 

This works fairly well for characters quantifiable by specif ic units (height, diameter, 
volume), but it is also used for more qualitative characteristics where somewhat subjective 
classes are taken as a basis. 

Some characteristics are determinant, regardless of the selection criterion. An 
example is resistance to pests or diseases: usually any trace of attack automatically 
eliminates the tree that is being assessed. 


A tree which appears phenotypically desirable and which is to be the subject of 
evaluation is called a candidate or preselected tree. Once its characteristics have been 
assessed and it has been accepted for further use, it becomes a select or plus tree. 

One its superior genetic characteristics have been ascertained, it is known as an 
elite tree. Several cycles of progeny trials are often required before a tree attains this 


1. Allard, R.W., 1964. Principles of Plant Breeding. John Wiley & Sons, Inc. New York. 

485 pp. 

2. Brown, C.L. & R.E. Goddard, 1961. Silvical considerations in the selection of plus 

phenotype. J. For 59: 420-426, 

3. Falconer, D.S., 1960. Introduction to Quantitative Genetics. The Ronald Press 

Company, New York, 365 pp. 

4. Gutierrez, V., M. and W.E. Ladrach, 1978. Iniciaci6n de un Programa de Mejoramiento 

Gene*tico de Cupressus lusitanica y Pinus patula en Colombia. Boletfn IFLAIC, 
No. 53: 3*19. 

5. Keiding, H., 1974. Selection of Individual Trees. In FAO/DANIDA Training Course on 

Forest Tree Improvement, Kenya, pp. 165-175. 

6. Ledig, F.T., 1974. An analysis of methods for the selection of trees from wild stands. 

For Sci 20: 2-16. 

7. Morgenstern, E.K. 1975. Review of the principles of plus tree selection. In Plus Tree 

Selection: Review and Outlook. Publication 1347, Canadian Forestry Service, 
Ottawa, Canada, pp. 1-27, 

- 128 - 

8. Smith, N., 1976. Selecci6n de irboles en Cachipo para estabiecer un huerto semillero 

^ e P* nus caribaea More let. II Seminar io Nacional de Plantaciones Fores tales 
S.V.I.F. MSrida, Venezuela. 42 p. 

9. Vollarreal, Raul, 1969. Consideraciones sobre un programa de Mejoramiento de 

especies Forestales en Mexico. Degree thesis. Escuela nacional de Agricultura, 
Chapingo, Mexico, 128 p. 

10. Wright, J.W., 1976. Introduction to Forest Genetics. Academic Press, New York 463 pp. 

Annex 1 


I- Straightness of trunk 

1. Straight 

2. Slightly crooked 

3. Very crooked 

II. Forking 

1. No forking 

2. Forked in upper third 

3. Forked in middle or lower third 

III. Defects (particularly for Pinus) 

1. No defects 

2. Recuperated visible defect 

3. Average to well-developed defect 

* v ' Flowering and fruiting 

1. Fruit 

2. Flowers only 

3. Neither flowers nor fruit 

The first three characteristics are determinant: the order of preference is: 
111, 211, 112, 121, 122, 221, 222. 

Class 3 will only be included in extreme situations, and would in any case 
indicate a fairly bad stand. 

Flowering and fruiting will serve as a basis for selection between trees in the same 
category, when the trees to be left or eliminated are marked. 

Data on d.b.h., height and any other obvious features will also be taken. 

- 129 - 

Annex 2 

A candidate tree and 4 neighbouring trees to be used for comparison are located. 

1. Vigour (basal area or volume) 

3 points will be awarded for each 10 percent of superiority of the candidate over 
the mean of the neighbouring trees, up to a maximum of 30 points (100% superiority). 
The same amount will be deducted if the candidate is inferior to the mean. 

2. Trunk 

Categories: 1. Straight 

2. Slightly crooked 

3. Very crooked (only for neighbours) 

1 point will be awarded if the candidate is in the best category and 1 additional 
point for each 0.25 of superiority over the mean of the neighbours. 
The same amount will be deducted for inferiority to the mean. 
Maximum number of points: 9. 

3. Quantity of verticils 

Points will be awarded as follows: 

1 point if the candidate is better than the worst of the neighbours but is below the 

2 points if it is equal to the mean. 

3 points if it is better than the mean but is inferior to the best neighbour. 

4 points if it is equal to the best neighbour. 

6 points if it is better than the best neighbour. 
Maximum number of points: 6. 

4. Average number of branches in the first 3 verticils 
The same as for total number of verticils. 
Maximum number of points: 6. 

5. Crown characteristics 

Categories: 1. Live branches only in upper third. 

2. Live branches as far as middle third. 

3. Live branches as far as lower third. 
Scoring as for trunk 

Maximum number of points: 9. 

- 130 - 

6. Diameter of branches 

The thickest branch (live or dead) up to the middle of the trunk will be taken. 

Categories: 1. No branches 

2. Thin branches (up to 1/10 of d.b.h.) 

3. Medium branches (up to 1/4 of d.b.h.) 

4. Thick branches (more than 1/4 of d.b.h.) 
Scoring as for trunk 

Maximum numbers of points: 13. 

7. Angle of branches 

Categories: 1. 90 or more with respect to the trunk above the branch 

2. Between 90 and 45 

3. Less than 45 
Scoring as for trunk 
Maximum number of points: 9. 

8. Attacks 

- 5 points:: healthy tree in a stand that shows slight evidence of attacks 

- 10 points: healthy tree in a stand considerably infested (more than 30%) 
Basic conditions 

1. A candidate tree must not be worse than the worst of its neighbours in any 

2. A candidate tree must not be below the mean of its neighbours in more than 2 
characteristics (negative values). 

3. No tree showing signs of attack by insects or diseases will be considered for 

Contribution of characteristics 
Maximum value of plus tree - 92 POINTS 

Vigour 32,67. Crown 9.8% 

Trunk 9.8% Diameter of branches 14.1% 

Verticils 6.5% Angle of branches 9.8% 

Branches 6.5% Attacks 10.9% 

- 131 - 



BJerne Ditlevsen 
National Forestry Service, Denmark 


Introduction 131 

Quantitative genetics 131 

Genotypic and phenotypic values 132 

Mean effect of a gene and reproductive value 133 

Deviations due to dominance and epistasis 133 

Variation of quantitative characters ... ... 

Concordance "between related individuals 

Heritability 136 

General and specific combining ability 137 

Selection and genetic gains 138 

Bibliography 139 


Quantitative genetics deals with the hereditary transmission of those differences 
among individuals that may be called quantitative, or differences of degree, unlike 
Mendelian genetics, which deals with differences of a qualitative nature (Strickberger, 

The basic genetic functions are the same in quantitative and Mendelian genetics, 
but in quairtitative genetics the differences are the result of differences of genes at 
many loci, while differences in Mendelian genetics are due to differences in genes at 
one or more loci, the different types appearing- in specific proportion (Falconer, 196*0. 


Most of the characters which are of interest in connection with forest tree impro- 
vement are controlled by a number of additive effects, genes, i.e. such typical quantitative 
characters as height, diameter, shape of the trunk and volumetric weight. Studies of 
the heredity of these characters must be made on populations, not individuals (Wright, 

- 132 - 

Analyses and studies of quantitative characters can employ a number of different 
statistical methods. 

Analyses of quantitative characters study the concordances between related indivi- 
duals, and one purpose of the analysis is to forecast the effect of a particular selection 
and to consider how it should be done. 

As mentioned above, statistical methods are essential, in studies of quantitative 
characters as are such magnitudes as average and variances. 

One important task will be to break down the phenotypic variation ipto its causal 
components suchi as external effects and genetic effects. It is also important to break 
down the genetic component into additive, dominant and epistatic components (Falconer, 
196U ; Shepherd, 1977). 


The genetic composition of a population is often expressed in the form of gene 
and genotype frequencies. To find a relation between gene frequencies and the quantitative 
differences evident in measurable characters, we may introduce a new term, the term "value", 
expressed in the units in which the character is measured (Falconer, 196U). 

The following terms are used: 

P * The phenotypic value of a given character. 

G = The genotypic value. "Genotype" means the complete genetic make up of an 

E * The deviation due to environment, or to all the nongenetic circumstances 

influencing the phenotypic value. 

The relation among the values may be expressed as: P g G + E. 

For the population as a whole, we have: 
The mean value m (E) or 

m (P) 

m (G) 

The above may be illustrated by a simple example in which the characters are deter- 
mined by a single locus with two alleles Al and A2. The individual genotype is given 
a value of -a, d and a respectively, as shown in the following equation: 





The genotypic value 

d * degree of dominance 

In the above example, the heterozygote A1A2 does not appear as the mean value of 
A1A1 and A2A2, but as the value d, which is called degree of dominance. 

If we imagine that the genes Al and A2 appear in a balanced population with the 
frequencies of p and q respectively, the genotypic value of the population may be calcu- 
lated as in the following table: 


f Genotype 

Frequency Value 

Mean value 


p -fa 

P 2 .a 


2.p.q d 

2pq . d 


q -a 

q . (-a) 



a(p-q) + 2pqd 

- 133 - 

The value a(p-q) + 2dpq is the genotypic value of the population insofar as it 
refers to this locus only, and at the same time the phenotypic value, since the sum of 
population deviations due to environment is zero. 

The mean of the population insofar as it refers to all those loci influencing the 
character in question (a quantitative polygenic character) is: 

M = I a(p-q) + 2 pqd 
loci loci 

Mean effect of a gene and reproductive value 

In the study of the transfer of a "value" of the parents to their descendants, 
we cannot use the genotypic value alone, since the parents transmit not their genotype 
but their genes to the next generation; instead, the term "mean effect" is used. It is 
defined as follows: the mean deviation from the average of the population obtained by 
individuals receiving the gene in question (assuming that the allele gene is taken at 
random from the population). 

The mean effect will therefore depend both on the gene in question and on the 
population in question. 

Based on the term "mean effect", the reproductive value (A) of an individual may 
be defined as follows: 

1. The sum of the mean effects of the genes (the genes controlling the character 
in question). 

The reproductive value is, by virtue of the definition quoted, a theoretical 
magnitude. Another, more practical, definition is the following: 

2. Twice the mean value of the progeny (measured in deviations from the average of 
the population), when the individual has paired at random with a series of 
members of the population. 

Reproductive value is also called additive genotype. 

In a balanced population, we find that the mean of the reproductive values of the 
individuals is zero. 

The reproductive value of an individual in a population may be calculated on the 
basis of the mean value of its offspring after free pollination. According to defini- 
tion number 2, the reproductive value may be calculated as follows: 

A = 2 x (average of population - average of progeny). 

The term "reproductive value", as an expression of that part of the hereditary 
mass that can be transmitted in an additive way to future generations, is very important 
in forest tree improvement. 

Deviations due to dominance and epistasis 

In the preceding paragraph on the individual, we mentioned the portion of geno- 
typic value that can be transmitted in an additive way to progeny. The remaining portion 
of the value of the individual consists of a non-additive gene effect between .the two alleles 
of a locus, and is called dominance (D). We therefore have: 

- 131* - 

For a balanced population we find that the mean of deviation due to dominance is 
zero, or 

m (D) 

If we study a polygenic character, there may also be effects of non-additive 
genes among loci. This phenomenon is called epistasis (I) and means that the genotypic 
value of polygenic characters may be broken down as follows: 

A + D + I. 


In the previous paragraph, we discussed characters and values in individuals; 
we now discuss the appearance of characters and values in the population. To this end, 
let us study first of all the variation of the characters and values and the relation 
between them* 

The values of the individuals and the variances of the population (in the form 
of variance components) may be set out as follows: 

P * G + E V p = V r + V_ 
i P iG E 

E 'V* V D* V I* V E 

VP Phenotypic variance 

VG = Genotypic variance 

VA = Additive variance 

VD = Dominant variance 

VI = Epistatic variance 

VE - Variance due to the environment 

The simple breakdown of phenotypic variance into the above-mentioned components 
of variance is based on the assumption that G and E axe independent of each other and 
that A,D and I are independent of one another, 

Variances can be used to express the relative importance of a cause of variance 
such as the relation between the variance components in question and the phenotypic 
variance. The importance of the genotype can, for example, be expressed as the ratio 

The components of genotypic variance and variance due to the environment cannot 
be determined directly from observation of a natural population. On the other hand, 
the above-mentioned components can be determined by trials or in experimental popula- 
tions when the parents are known. 

Concordance between related individuals 

The components of variance in the previous paragraph can be traced to variations 
and therefore we call these components causal variance components. 

On the contrary, if we wish to measure the degree of relationship between rela- 
ted individuals, another distinct division must be made in the phenotypic variance, the 
grouping of individuals into families. These components can be directly estimated on 
the basis of phenotypic values, and therefore we call them observed variance components. 
To avoid confusion, we will use V for the causal components, <r for the observed 

We discuss below the four most usual kinds of relationship, which are: 

1. The progeny and one of the parents (the other taken at random from the popula- 

2. The progeny and both known parents; 

- 135 - 

3- Half-sibs 
U.-Full sibs. 

The concordance between parents and their offspring can be indicated by the incli- 
nation (b) of the regression of parent-children: 

Offspring (0) 

coefficient of inclination 

Parents (P) 

b (coefficient of inclination) is estimated as: 

P = parents SP sum of pruducts 

= offspring SS sum of squares 

Concordance among siblings can be indicated by size: 

2 2 

or or _, Variance among groups of siblings 

_ & B 

o* B + a, cr s Variance among individuals within groups of 


Now we want to link the observed components with the causal components, i.e. to 
find the ratio between the observed values of b and t and the causal components VA and VP. 

It is possible to deduce these ratios for the above mentioned relationships. We 
give one example below; the relationship of progeny and one of the known parents. 

The concordance is indicated by the coefficient b OV PO 

According to definition 2 of the reproductive value of an individual the mean (m) 
progeny is equal to 1/2A, 

or the co-variance must be calculated between one of the parents (with the value - G) 
and the mean of the progeny (with the. value 1/2 A). 

Cov po = Cov(G, 1/2 A) Cov(A +D t 1/2 A) 

SP po ((A + D) (1/2 A) = 1/2$AD + 1/2(A* 

Assuming that A and D are indi pendent, the sum of their product is A . D 0. 


136 - 

The genetic interpretation of the other relationship may also be deduced. The 
results are summarized in the following table: 

Relationships Concordance observed Genetic interpretation 

Offspring and one b 1/2 


Offspring and the b ~j 

two known parents v ^ 

t W 

PMll sibs t 1/2VA^1AVD 


Hhe concept of heritability of a character means the proportion of total variance 
which is due to the action of genes. 

Two concepts of heritability are used, which are defined as follows: 

2 VG 

1. Heritability in its widest sense h b g * ^ 

2 VA 

2. Heritability in its narrower sense h * =: 

n. s v* 

It is important to note that heritability is not any fixed magnitude, but depends 
both on the population and on the environment. Indication of heritability should always, 
therefore, be accompanied by an indication of the environment of the components. 

As a general rule, we may say that the herit abilities of current stands are relati- 
vely low, due to the fact that the environment has a relatively great impact on the pheno- 
typic value of the individual. However, heritability can be fairly high in well-planned 

As mentioned above, a genetic interpretation may be made of the coefficients of 
regression observed (parent -off spring) or coefficients between classes (between siblings). 
Likewise, the coefficients observed can be used as estimates of heritability. We give 
below a summary of the relations between the coefficients and h2 n.s. in the four fol- 
lowing relationships: 


1. Offspring and one parent b * 1/2 h 


2. Offspring and both parents b = h 

3. Half-sibs t * lA h 2 
k. Pull sibs t 1/2 h 2 

normally, heritability is linked to individuals, but both family heritability and 
heritability within the family can be estimated from the guidelines described at the 
beginning of this section. Family herit abilities are higher than the corresponding 
individual herit abilities since, using average values (average of family), an important 
reduction may be achieved in the environmental effect component. 

We indicate below how heritability may be calculated on the basis of the results 
of a progeny test. The test includes groups of half-sibs and is arranged as a random 
block design. The components of variance described in the following diagram of variance 
analysis are used to make the calculation: 

- 137 - 

Source of variation 


variance components 

Progeny (d) 


^ ,4 

Blocks (b) 


o* *t" n. or 






With groups of half-sibs, the variance between the groups can be interpreted as 
lA VA, or: 

lA VA; VA 

VE =6 /r 

If ve assume that there is no dominance, the phenotypic variance could be written 
as follows : VF * VA + VE. 





The variance between families was formerly divided into genetic components called 
causal components. However, the variance between families can also be divided into the 
two following components, which can be observed: 

1. General combining ability g. c. a. 

2. Specific combining ability s. c. a. 

The g.c.a. of an individual is defined as the average offspring resulting from 
the crossing with various other individuals. The s.c.a. of two specific individuals is 
defined as the deviation of their offspring from the average g.c.a. of the two indivi- 

The average offspring between two individuals x and y can therefore be described 
as follows: 


Consequently, the variance between crossings may be analysed in two components: 
variance in the general combination capacity and variance in the specific combination 
capacity. The latter appears in the statistical analysis as the interaction component. 

From definition 2 of reproductive value it may be seen that the g.c.a. is equal 
to half the reproductive value of the individual. 

The specific combination capacity depends exclusively on non-additive genetic 

A knowledge of g.c.a. and s.c.a. is important for the selection and combination of 
material for future seed production. 

G.c.a. and s.c.a. can be calculated on the basis of crossing designs that will be 
described in more detail in a later article, and we give below only one example of the 
calculation of g.c.a. and s.c.a. 


1. Mean values of four progenies: 



















2. Deviation from the average total of the progeny and calculation of the g.c.a. 

















3. Calculation of the s.c.a. through the equation 
B.c.a. 13 * -30 -(-25) - (-10) * 5 



10 - 25 - (-10) * -5 

-20 -(-25) - 10 = -5 

1*0 - 25 - 10 * 5 


Some of the above mentioned genetic parameters, particularly heritability, are 
very important for selection and the estimation of genetic gains through selection. 

The strength of selection in a population may be expressed as the selection diffe- 
rential S which indicates the difference between the mean of the population and the mean 
of the selectedgfraction, or as intensity of selection i ^hich is expressed through the 
* *- - being equal to the phenotypic dispersion of the population. 

formula i 

If we select the best individuals in a population, we can express the genetic 
gains (R) as the displacement of the mean among the progeny of the selected individuals 
and the entire population. 

The relation is illustrated in the following figure: 

- 139 - 

Offspring (0) 

s b = coefficient of slope 

Average of the tvo 
parents (P) 

The slope of regression which expresses the relation between S and R is equal to 

P g p 

The coefficient of regression b = h -r-, hence R = h S. 

n. s R n.s. 

The selection differential S. depends on the dispersion of the population, and it is 
therefore a good idea to standardize it by dividing it by phenotypic dispersion, as in 
the case of i. 

Using i instead of S the following function is obtained: 

R = i.hj ap 


Falconer, D.S. 

Shepherd, K.R. 

Strickberger, M.W. 

Wright, J.W. 

Introduction to Quantitative Genetics, Oliver and Boyd, Edinburgh 

General Genetics and Inheritance, International Training Course in 
Forest Tree Breeding, Canberra, Australia. 

Genetics, The Macmillan Company, New York, U.S.A. 

Introduction to Forest Genetics, Academic Press, New York, U.S.A. 

- 140 - 

Marcelino Quijada R. 

Institute de Silvicultura 
Unlversidad de Los Andes 
Her Ida, Venezuela 



Definition 140 

Practical use 140 

General terminology 141 

Methods 141 

Cuttings 141 

Grafting 142 

Layering 143 

Tissue cultures 143 

Incompatibility 143 

Bibliography 145 


Plants can be propagated sexually or asexually. The basic difference is the pro- * 
cess of fertilization that occurs in the first but not in the second method. 

Asexual propagation, in turn, takes two forms: apomixis and vegetative propagation. 
Apomixis is propagation through the development of one of the gametes (the ovule) without 
fertilization, or of a cell without reduction, to form seeds or seed-like organs. 
Vegetative propagation is propagation from well-differentiated vegetative parts. 


Practical use of vegetative propagation methods is based on two biological consi- 

a) Maintenance of the physiological condition of the parent tree in the propagated 

b) Maintenance of genetic constancy. That is to say, the part propagated is gene- 
tically identical to the original individual. 

Vegetative propagation has been widely used in breeding for, among other things: 
a) the establishment of clonal seed orchards; 

- 141 - 

b) the establishment of clone banks, in which controlled pollination is effected, 
owing to the possibility it affords for obtaining flowers at low height; 

c) the propagation of special breeding material: exceptional hybrids (e.g. hetero- 
tics) that are lost through sexual reproduction, sterile hybrids, etc.; 

d) the propagation of selected plants on a large scale. 
Its usefulness depends on, among other factors: 

a) The ease with which the species can be manipulated. Many species are difficult 
to propagate vegetatively, others are extremely easy. This often affects 
production costs, both for the establishment of orchards and for large-scale 
production of plantation material. 

b) The extent to which development of the parts propagated can be controlled. 
In some cases the phenomenon of topophysis occurs: the development of the 
propagated part is influenced by the part of the tree from which it comes, 
e.g. lateral branches sometimes tend to grow in a horizontal direction. 

Another phenomenon which affects development is graft incompatibility: in grafting 
the scion may be rejected, sometimes after a year or more. 


The original tree from which the parts to be propagated are taken is known as the 
ortet. Each part already propagated is a ramet . A set of ramets from the same ortet is a 
clone. Differences observed between trees propagated by the same method is called clonal 


There are three main methods of vegetative propagation: cuttings, layering and 

Cuttings are sections taken from the tree and put to root in an appropriate 

Grafts are plants obtained by fusing a part from the tree to be propagated with 
another part which has its own root. 

Layers are sections of the tree which are induced to root and then separated 
from the tree. 

For breeding purposes the methods most used are cuttings and grafts. Layering 
can be of help in special cases as a transitory measure. 


Cuttings may be of wood or leaves. The first come from branches or the trunk and 
are more important in forestry. Leaf cuttings are more common for ornamental plants 
particularly those with fleshy leaves. 

It is the most economical method of vegetative propagation in species with good 
reserves of aqueous tissue, such as Bombacaceae. 

Rooting depends on, among other things, the age of the plant, the condition of 
the cutting, the time of collection, the rooting medium and special treatments. 

- 142 - 

Tree age is difficult to determine in natural tropical forests, buy the dimensions 
of the tree, in particular the diameter, can be used as a guide. Very stout trees, more 
than 50 cm. in diameter, are usually old and therefore do not root easily; when rooting 
is achieved, the root system often develops poorly. The best propagation is achieved 
with young trees, but this often runs counter to specific purposes, such as the production 
of flowers, fruit and seeds. 

As regards the condition of the cutting, very woody cuttings do not sprout or 
root easily. Very herbaceous cuttings tend to be very susceptible to drying out. An 
intermediate condition is therefore best, particularly when the cutting is planted out 
in the field immediately. A guide for the collection period is the stage of activity 
of the plant; the most favourable period is when the buds on the tree are starting their 
most active growth. In the low-lying tropical zones of Venezuela this coincides with the 
onset of the rainy season (March/April). 

The rooting medium must guarantee sufficient but not excessive moisture; this is 
normally achieved with a medium, semi-sandy textured soil and adequate atmospheric humi- 
dity. This is easy to control under artificial conditions, but in the field it can only 
be guaranteed during the rainy season. 

As regards treatments, rooting hormones such as indoleacetic, indolebutyric and 
naphtalenacetic acid are useful for small-scale work (and even for orchards). For com- 
mercial purposes, powdered dusts such as Root one, Hormodin, Hormonagro, etc. or pure 
solutions of 100 to 1 000 p. p.m. are recommended. 

After the cutting has been planted, the upper surface, if it is the site of the 
cut, should be protected with a substance to reduce evapotranspiration. 


This involves the removal of a vegetative part from the parent tree (the scion) 
and its attachment to a part with its own root (the stock) so that the tissues fuse. 

Depending on the taxonomic relations between stock and scion, we have hetero- 
plastic grafts if they belong to different species (e.g. Cedrela odorata on . angusti- 
folia) and homoplastic grafts if they belong to the same species. Homoplastic grafts 
are called autoplastic if stock and scion have the same genotype. 

According to the position of the scion, we have top and side grafts, depending 
on whether the scion is inserted into the top or the side of the stock. 

Top-grafting entails beheading the stock, which in many cases results in a dras- 
tic reduction in the plant's foliage. This technique is used for many hardwood species. 

Side grafting makes it possible to maintain adequate foliage on the stock, and 
has been much used for conifers, where the moisture factor is very important. 

Diagrams of some grafts commonly used for forest trees are given in Annex 1. A 
good, step-by-step description of grafting and layering with special reference to pines, 
is given in Dorman (1976). 

The cuts should be made with grafting knives, which should be kept clean and 
sharp, in order to ensure a clean cut which will not be a focus for infection. These 
knives also have a metal or hard rubber spur to separate the bark in budding. 

Cutting must be rapid and uniform for a clean, even cut. 

The scion should be bound to the stock with a strong but slightly flexible mate- 
rial. A semi-elastic plastic strip called grafting tape is usually used. It may also 
contain a fungicidal compound to prevent attacks by fungi. 

- 143 - 

Outdoor grafts must be effectively protected to prevent the scion drying out. A 
double covering is often used, the inside layer consists of a plastic bag which helps 
to retain moisture, and the outer of cloth or non-transparent paper which protects the 
graft from the direct rays of the sun. As a precaution, it is advisable to effect the 
graft on a cloudy, slightly humid day. 

In selecting the scion the main consideration is its stage of growth. Buds that 
are in full activity should be sought; this can be checked by their size and by the time 
of collection (little differentiation and collection at the onset of vegetative growth). 

The scion can be the bud itself with part of the bark to attach it to the stock, 
or it can be a piece of the branch with one or more buds. In the latter case the piece 
of branch must be neither too soft nor too hard, a second-year growth is ideal. 

Care must also be taken to see that the scion is kept in a cool, damp place for 
no more than two days, if not used immediately. 

The stock must be young and healthy. The main guide is compatible thickness of 
scion and stock. Diameters of between 0.5 and 2 cm. are preferable to ensure good adhesion. 


It may often be difficult to propagate a species by cuttings, and grafting may 
present problems. One possible solution is layering. The procedure consists of making 
a wound in a section of the tree and covering this with a medium which retains moisture 
(moss, earth, etc.). Healing causes the formation of a callus from which adventitious 
roots may grow. 

This process is helped along by hormone compounds similar to those used for cut- 

The most common type of layering is aerial layering in which the covering is a 
medium other than soil and the operation takes place rather high off the ground. When 
long, low, flexible branches are available, these can be inserted into the soil; we then 
have ground layering. Usually the part propagated is a branch, but in some species (e.g. 
Platymscium sp.) interference with the roots can cause them to sprout, constituting a 
form of layering in that they do not separate from the stem. 

The recommended size is as for cuttings; branches with a diameter of 1 to 2 cm. 

One of the biggest problems of this method is transplanting, which often results 
in considerable losses. It is recommended that the first transplant be made in a fairly 
loose (sandy) medium to encourage root adaptation. The plant can then be transplanted 
into a more compact medium. 

Tissue culture 

Another form of vegetative propagation consists of tissue cultures, (cultures of 
cells with the potential mitotic activity, in an appropriate medium under aseptic condi- 
tions). The method has been tried with varying degrees of success in gymnosperms and 
angiosperms, and everything seems to indicate it is feasible for many forest species. 
The great advantage is that only a very small portion of material is needed. The main 
disadvantage is the very controlled work conditions required and the same limitations as 
for cuttings, grafts, etc. (e.g. genotyplc variation in ease of propagation). 


One of the main problems in grafting is incompatibility which produces rejection of 
the scion by the stock. 

Rejection simply indicates (of ten) that the cut areas have not been fitted together 
well (e.g. uneven cuts minimize points of contact). Sometimes, however, rejection is 

- 144 - 

an indication of conditions inherent in genetic somatic differences between the tissues. 
This phenomenum may occur at an early stage, making it necessary to effect more grafts 
to fulfill a given quota (e.g. clone(s), problems with clone(s)). A greater problem is 
that of delayed incompatibility which may emerge one or more years after the grafts have 
been established in the field, thus leading to greater losses, e.g. losses of producti- 
vity in the case of seed orchards. 

In many species external signs have been detected which indicate incompatibility. 
These indlude: a) consistent early defects in specific trees, b) unequal rate of growth 
of scion and stock, c) excessive growth in, above or below the juncture zone, and d) 
abnormalities in colouring and leaf development of the scion. 

The last three are obvious signs of delayed incompatibility which appear between 
six months to ten years after grafting. 

This phenomenon may occur not only in heteroplastic grafts, where it might be 
expected because of taxonomic differences, but also in homoplastic grafts, i.e. within 
the same species. In the latter instance, it has more to do with differences in prove- 
nance or geographic source. One way of controlling this phenomenon is to seek the great- 
est possible affinity between the plant serving as stock and the tree from which the 
scione are taken, to ensure optimum histological affinity. 

It is to be noted that there is no best method of grafting: success often depends 
on the skill of the grafter using a specific method. Various tests have shown different 
degrees of difficulty in manipulating scion and stock, but the final results have been 
practically the same as regards percentage of take, as shown below. 

Vegetative propagation characteristics of some forest species in Venezuela^ 
Species Ease of vegetative propagation 

Bombacopsis qulnata VE 

Cedrela odorata M - E 

Tabebuia rosea M - E 

Swietenia macrophylla D 

Anacardium excels urn M - D 

Pithecelobium saman yp 

Podocarpus rospigliosil D - VD 

Cordla alliodora M - D 

Cordia apurensis M - D 

Podocarpus oleifolius M 

Hura crepltans M - E 

Tectona grandis 

Gmelina arborea 

Plnus caribaea v. hondureneis M - E 

Plnus oocarpa M - E 

Pinus radiata M - E 

Plnus patula M - E 

- 145 - 

N.B. VE Very easy, percentage above 80% 
E - Easy, average percentages 60% 
M Moderately difficult, percentages around 40% 
D - Difficult, percentages around 20% 
VD - Very difficult, percentages usually 0% 
Species between 2 classes show great clonal variability. 


Dorman, K.W. The Genetics and breeding of Southern Pines. U.S. Dept. Agriculture, 

1976 Agriculture Handbook no. 471. Washington D.C. 

Hartman, H. and D. Kester. Propagacion de Plantas. Companla Editorial Continental, 
1967 S.A. , Mexico. 693 pp. 

Jett, J.b. Vegetative Propagation. 

1969 In Forest Tree Improvement Training Course. 

N.C. State University, Raleigh N.C. pp. 217-225 

Jett, J.B. Incompatibilty. In Forest Tree Improvement Training Course, 
1969 N.C. State University, Raleigh, N.C. pp. 226-230 

Koenig, A. and Melchior, G.H. Propagacion Vegetativa en Arboles Fores tales. Proyecto 
1978 Investigaciones y Desarrollo Industrial Forestales - COL/74/005. PIF 
no. 9. Bogota, Colombia. 

New Zealand Forest Service. Special Issue on Vegetative Propagation. 
1974 N.Z. J. For. Sc. .4(2): 119 - 458. 

Quijada, M. and V. Gutierrez. Estudio sobre la propagacion vegetativa de Especies 
1972 Forestales Venezolanas. Rev. For. Ven. 21: 43-56. 

The Swedish University of Agricultural Science. Vegetative Propagation of Forest 

1977 Trees - Physiology and Practice. Uppsala, Sweden, 159 pp. 

Wright, J.W. Introduction to Forest Genetics. Academic Press, New York, 
1976 463 pp. 

- 146 - 




In a graft the cambium of the stock and the scion should coincide at least on one 
side in order to form the new tissues uniting the two parts rapidly. The cuts must he 
flat and smooth so that as few cells as possible are damaged. To obtain these conditions 
various grafting methods have been developed, depending on the diameter of scion and 
stock. The main grafting methods used for forest trees are the following: 

a) Cleft graft 


Bark of scion 

Stock |.fifc'f . Bark of stock 


In cleft-grafting the scion cut in the form of a wedge is inserted into a cut made 
in the top of the stump. Scion and stock must be of approximately the same diameter. In 
simple or radial cleft-grafting, the scion occupies only one side and may have a diameter 
inferior to that of the stock. 

b) Splice or whip and tongue graft 





Scion and Stock (similar diameters) are joined by means of a long diagonal cut so 
that the cut on the stock coincides with the cut on the scion. 

J7 From Koenig and Melchior (1978). 

c) Veneer-graft 

- 147 - 


Stock Scion 

This method is used when the diameters of stock and scion differ considerably. The 
stock is prepared by a slanting cut which penetrates as far as part of the xylem, and the 
cut piece removed. The scion is prepared so that the cut coincides with that in the stock. 

d) Underbark graft 



This method Is also used for different diameters, only the bark of the stock being 
cut. The scion, cut in a triangular shape, is placed in the opening. 

b) Budding 

Bud shield 

The insertion of buds used in the propagation of fruit trees and roses, called 
budding, can also be used for tropical forest trees, if only a small number of scions are 
available, or if the methods described above do not prove successful enough. 

- 148 - 


B. Ditlevsen 
National Forestry Service, Denmark 



Introduction 148 

Purpose of controlled crossings 148 

Selfing 149 

Crossing systems with unknown father 1 50 

Free loss of flowers I 50 

Poly-crossing 150 

Crossing systems with known father 151 

Complete diallel crossing plan 152 

Modified diallel crossing plan 152 

Partial diallel crossing plan 154 

Factorial crossing plan 1 55 

Single pair pairings 157 

Choice of crossing plan 157 

Bibliography / 159 


One of the most important decisions to be made in those genetic improvement programmes 
that include controlled crossings, is the choice of a crossing system. Apart from tree 
selection, the choice of a crossing system is the only way in which the tree breeder can 
improve a genetically variable population. 

We describe below a series of important aspects to consider when choosing a crossing 
system. In practice, however, it is often difficult to choose the best system, since there 
are usually different competing requirements that cannot always be satisfactorily met by 
one 'system alone. The final choice, therefore, will probably be a compromise solution. 
Lastly, there may be some practical and economic constraints. 


As indicated above, the purposes of controlled crossings are many and various, and 
we set out below the most important ones for forest tree improvement (Brown, 1977, Roberts, 

1. Determination of general combining ability (g.c.a.). It is usually difficult to 
estimate the reproductive value, or general combining ability of an individual 
tree, from the phenotypic value of the individual; only studies of its progeny 
will provide us with the necessary information on the capacity of the individual 
to transmit its good characters to its offspring. 

- 149 - 

Information on the g.c.a. of individuals can in practice be used for: 

a. Choosing the best individuals for seed orchards; 

b. Genetic thinning in existing seed orchards; 

c. Choosing the parents of the progeny which it is desired to use in future 
genetic improvement. 

2. Determination of specific combining ability (s.c.a.). Data on specific 
combining ability can be used in establishing seed orchards of two 
clones; this makes it possible to select for special effect in the 
offspring of two individuals. 

3. Determination of variance in general and specific combining ability* 

4. Estimation of heritability. Knowledge of both the variance in 
g.c.a. and s.c.a. and heritability is very important in deciding 
for the establishment of the best procedure for genetic improvement 
and for optimization of, for example, the number of individuals for 
progeny and the number of progeny per clone. 

5. Production of material for selection of individuals for the next 

feneration of seed orchards. We may assume that the best individual s 
n a group of siblings are better than the average of the parents and, 
consequently, it is better to select the best individuals within the 
best groups of siblings for use in the future seed orchard. 

6. Production of material for continued improvement. To raise the 
genetic quality still further from the level achieved by a seed 
orchard of clones whose progeny has been evaluated, it is necessary 
to recombine the hereditary characters through crossings and new 
selections. This procedure can be repeated several times before selecting 
the individuals that will form part of the future generations of seed 
orchards. In this way it is possible to eventually achieve several 
combinations of genes that are very seldom found growing wild but that offer 
great advantages from the breeding standpoint. 

7. Estimation of genetic gains in the first generation of seed orchards and 
in successive generations. Decision-makers should be able to get 
information on the degree of improvement (genetic gains) achieved in seed 
orchards as compared with seeds from stands, and on the genetic gains that 

can be expected from the establishment of seed orchards of the next generation. 

Lastly controlled crossings could be useful in practical forestry work to obtain 
plus material. For example, two clones might have a very high capacity for specific 
combination, but without simultaneous flowering periods; they could not therefore, 
be used in a seed orchard of two naturally pollinated clones. 

The establishment of controlled crossings of forest trees is an expensive 
time-consuming operation. In consequence, it is often best to use crossing designs 
that meet several of the desired objectives at the same time. 


Controlled self-pollination is studied mainly in research on selfing. In 
seed orchards, selfing clones can produce much selfed seed, which in turn produces 
minus plants* 

Controlled self-pollination can also form part of an intra-crossing/crossingg 
progranme in which attempts are made to produce strongly intra-crossed individuals 
which are later crossed with other, also intra-crossed, individuals to achieve a 
heterosis effect. The method is better known in agriculture, but it has also been 
tested in forest tree improvement (for example, with Larix, see Keiding (1968). The 
possibilities of using intra-crossings and crossing* in forest tree breeding are 
discussed by Lindgren (1975). 

- 150 - 


In this paragraph we shall discuss two designs: free loss of flowers and 

Free loss of flowers 

After free loss of flowers the seed is usually collected by one of the two 
following methods. 

1. Collection of seed from plus trees selected from a stand. The plus tree 
has been fertilized by the other trees in the stand, and it may also have 
many selfed seeds. 

2. Collection of seed from established seed orchards. If the seed orchard or 
the plantation is producing large quantities of pollen, it may be assumed that 
individual clones in the plantation have been fertilized by the other clones 
in the plantation or, in some cases, by self-pollination. 

In both cases, the uncertainty of the relationship between the individual seeds 
is considerable. In analyses of results of tests and in later interpretation it is 
normal to assume that individuals of the progeny after free loss of flowers are half- 
sibs t or that they have the same mother but different fathers. 

We have already mentioned that in certain cases certain quantities of selfed 
seed may be expected, and also, it may be assumed that some of the progeny are full 
sibs (that they have a common mother and father). 

The uncertainty as regards relationship may lead to erroneous interpretations, 
and the most important disadvantage of free loss of flowers is the difficulty in 
determining the general combining ability due to the fact that the fathers are unknown 
and that they vary. 


The disadvantages of free loss of flowers, namely the risk of self-pollination 
and fertilization by different pollen for different mothers, can be avoided through the 
so-called poly-crossing system. In poly-crossing, artificial pollination is effected, 
using a mixture of pollen from a large number of fathers (not, however, from the mother's 
clone). In this way self-pollination may be avoided, and all the mothers are pollinated 
with a constant mixture of pollen. 

So many different fathers are involved that specific effects of combination, if 
any, are neutralized, and poly-crossing is therefore a good, cheap system for determining 
g.c.a. However, it should be mentioned that non-randomized pollination (because the 
pollen of some clones is more viable than that of others) can result in erroneous estimates 
of g.c.a. (Roberds, 1969). 

One characteristic common to free loss of flowers and poly-crossing is that, 
since the fathers of the offspring are not known, the possibilities of making a selection 
of progeny for use in the next generation of seed seedlings are considerably limited. 
On the one hand, it is impossible to select progeny with a good quality father, and 
on the other hand, there is the risk of choosing individuals already related to each 

Systems in which the fathers are unknown do not allow for evaluation of the possible 
specific effects of combination. 

Nevertheless, the above-mentioned disadvantages of the unknown fathers system 
are offset by the fact that these systems are cheap and simple to implement, and both 
the trials and the statistical analysis are easy to carry out. 

The following is a diagram of the analysis and interpretation of a poly-crossing 
design laid out in a field trial as a random block design. The genetic interpretation 

- 151 - 

has been based on the assumption that the progeny are half-sibs. 
Variance Analysis 

Source of variation 



expected MS 





M l 



(a-1) (b-1) 

M 2 


Within plots 


M 3 


Genetic interpretation 

a - 




V P - 

-- . 

v. ; 

Variance due to families 
Environmental variance among plots 

Environmental variance among trees in the plots 
Additive variance 

Dominant variance 

Environmental variance within plots 

Phenotypic variance 

Genotypic variance 


In this group of crossing systems the offspring are full sibs, i.e., the individual 
offspring have a common mother and father. Therefore, as a general rule, we may expect 
to have better information from progeny trials than from systems where the father is 
unknown. However, systems where both the mother and the father are controlled are 
more difficult, so various fairly complete designs have been developed; they are described 
In detail below. 

- 152 - 

Complete diallel crossing plan 
Figure 1 shows a complete diallel crossing plan. 
Fathers 12345678910 


V N 



































V s 

S, x 






















v x 














































v X 

Fig. 1. Complete diallel crossing plan. Self-pollination is indicated in the figure. 

This crossing plan is the best one, since it includes all crossing possibilities 
and provides the most complete information on the genetic characters of the clones 

This design can provide information on the general and specific effect of 
combination and on its variances. The material also constitutes the best starting point 
for the selection of superior individuals or pairs of clones suitable for two-clone 

For practical purposes this design is, unfortunately, very difficult to use. 
In the first place, it usually yields clones that produce too small a quantity of male 
or female flowers; however the most important disadvantage is an economic one. A complete 
diallel crossing plan with, for example, 20 clones, will require 400 controlled crossings, 
or 380 if self-pollination is avoided. However, it is not very realistic to spend 
so much on each clone, and a cheaper design, such as the one described below, is more 
generally used. 

Modified diallel crossing plan 

One way of shrinking the crossing plan, and thus making it cheaper, is to omit 
reciprocal crossings and self-pollinations, as shown in Figure 2. 

- 153 - 






X X 


X X X X 

X X X X X 

X X X X X X 

X X X X X X X 


Fig. 2. Modified diallel crossing plan. 

This design provides roughly the same information as the full diallel crossing 
plan, but the limited material in the plan does not, of course, guarantee the same accuracy 
in trials and determination of parameters. 

Below is a diagram of the statistical analysis and the genetic interpretation of 
a modified half diallel crossing plan in a random block design. 

Variance analysis 

Source of variation 



Expected MS 







M l 

a w * ^e * wb s * wb(n ~ 2)0 2 



M 2 

a 2 + wo 2 -i- wba 2 
we s 



M 3 

a 2 * wo 2 
w e 

Within plots 

nb(n-l) (w-l)/2 

M 4 

a 2 


GCA - General combining ability 
SCA - Specific combining ability 

Genetic Interpretation 

1/4 v_ 

1/4 V D , 

4 . a' 

- 154 - 

Partial diallel crossing plan 

This design may be distinguished from the complete and modified diallel plan in 
so far as a clone has not been crossed with all the other clones. The design can have 
a series of different conformations (Braaten, 1965) and in Figure 3 we indicate two 
different types. 




X X X X 




X X 





X X X X 




X X 




Disconnected partial diallel design. 













10 11 

x x 

X X 

X X 

X X 

X X 


X X 


X X 

Brown's partial diallel design (Kempt borne and Cur now, 1961) 
Fig. 3 Partial diallel crossing plans. 

As may be expected, these designs are less efficient than the full design and 
the modified diallel design. However, to make up for this, a large number of offspring 
can be tested at a fairly low cost. Unfortunately, missing values, which cannot be 
avoided in large-scale crossing plans, considerably complicate the work of calculation. 

As illustrated in Figure 3, there are various different designs in the group of 
partial diallel designs, and we give below an analysis of one of the most frequent, the 
disconnected partial diallel design. 

- 155 - 

Variance analysis 

Source of variation 



Expected MS 





Block x diallel 

(b-1) (d-1) 

Families within 


M l 

222 2 
w * WQ e * bwQ s * bw < n ~ 2)a 



M 2 

a 2 + wa 2 + bwa 2 
we s 


d( b -l,(c-l, 

M 3 

w e 

Within plots 


M 4 

a 2 


Genetic interpretation 

M. - M, 
9 bw(n-2) 

, 2 - M 2" M 3 
8 bw 
.2 M 3- M 4 


v^ - a 2 - a 
G g s 



rr 4- rr 

Factorial crossing plan 

In this design a number of mother clones are crossed with the same number of father 
clones. Often there is a small number of fathers, also called common testers. This 
design can also be considered a sort of full diallel design including all the combinations 
of one group of mothers and another group of fathers. See Figure 4. 

- 156 - 














































Fig. 4 Factorial crossing plan. 

The design is widely used in the United States under the name of North Carolina 
II. Normally four different fathers are used for the crossing plan, but as a general 
rale, the number of fathers in the plan must depend on the magnitude of the specific 
combining effects. If major specific combining effects are likely, the estimation of 
g.c.a. of the mothers may turn out to be highly erroneous if too few father trees are 

Since the design often includes very few fathers and since the same clone does 
not appear as mother and father, it is difficult to compare the general g.c.a.'s of each 
parent . 

It is also difficult to make a progeny selection for the next generation of seed 
orchards, since selected individuals, especially when only a few fathers are used, are 
very often already related to each other. 

One advantage of this design is that its implementation is simple and therefore 
relatively cheap; at the same time it makes analysis of the results easier. We give 
below a diagrammatic summary of a factoral design. 

Variance analysis 

Source of variation 



Expected MS 





M l 
M 2 

a w * ^e * ^^mf * bwf a m 

Fatherx x mothers 

,.-!, ,-U 

M 3 


Within plots 

<*-!) (b-1) 

M 4 


hBf (*-!) 

M 5 

a 2 


- 157 - 
Genetic interpretation 

- 1/4V A ' V A 

V A 

a 2 - M. - v_ - a 2 - a 2 - a 2 + v_ v ft a 2 + a 2 + a 2 . + a 2 + a 2 

w 5 Gm f mEw;P mfmfew 

Single pair pairings 

In this system each clone is involved only once as mother or father. 

The system works particularly well if the objective is to produce a population 
for selection of individuals for new seed orchards or for use in continuous breeding 

Another advantage of the use of single pair pairing is that a large number of 
clones may be tested under the same plan and normally it is very expensive to produce 
offspring on the basis of controlled crossing. 

On the other hand, the possibilities of estimating the g.c.a. and the variance 
of the general and specific combining abilities are normally not very good. If the 
specific combining effects are not important, a quick estimate of g.c.a. may, however, be 
made in order to set aside the worst of the clones tested. If, on the other hand, 
the specific combining effects are very important, this design may be used to select 
the best combinations of clones for use in two-clone plantations. 


The definitive choice of a crossing plan will depend mainly on the purpose of 
the controlled crossings, as described in the first paragraph of this article. The 
advantages and disadvantages of the different crossing systems and designs have been 
briefly discussed in the paragraphs above, and we provide a summary of all the most 
important aspects in Table 1. Lindgren (1977) and Buijtenen (1976) have given a more 
detailed analysis of the different designs. 

Apart from the problems discussed, which are important in the choice of a crossing 
plan, a series of problems of a practical nature may arise in the implementation of 
large series of crossings; for example the differences between clones insofar as their 
time of flowering is concerned. 

We may also mention problems related to the transfer of the crossing plan to a 
field design. If possible, a simple, sturdy test design, such as the random block, 
should be chosen* With many different offspring it might, however, be necessary to 
use an incomplete design. See Braaten (1965). 

- 158 - 

Table 1. Comparison of crossing plans 


of CCA 

Selection of 
plus trees 


Determination of 
variance of GCA 
and SCA 

Free loss 
of flowers 


Possible, but 




Very good 

Very little and 
only if the 
depression of 

Very low 

Good determination 
of the variance of 

is reduced 




Very high 
where there is a 


large number of 




Very high 
where there is a 

Very good 

large number of 



Very good 


Good determination 
of the variance of 

GCA. The variance of 

SCA can be determinei 

but it is difficult 

from the point of 
view of data process* 




Only in a few 
cases and 



where intra- 

depression is 




Very low 



GCA - General combining ability; SCA - Specific combining ability 

- 159 - 


Brown, A.G. , 1977. The Strategy of Tree Improvement. International Training Course in 
Forest Tree Breeding, Canberra, Australia. 

Braaten, M.O., 1965. The Union of Partial Diallel Mating Designs and Incomplete Block 
Environmental Designs. Inst. of Statistics. Mimeograph Series No. 432. 

Keiding, H. , 1968. Preliminary Investigations of Inbreeding and Outcrossing in Larch. 
Silvae Genetica 18(5). 

Kempthorne, 0. and Curnow, R.N., 1961. The Partial Diallel Cross, Biometrics 17. 

Lindgren, D. , 1975. Use of Selfed Material in Forest Tree Improvement. Dept. of Forest 
Genetics, Research Notes 15. 

Lundgren, D. , 1977. Genetic Gain by Progeny Testing as a Function of Mating Design and 
Cost. Third World Consultation on Forest Tree Breeding, Canberra, Australia. 

Ro herds, J.H., 1969. Progeny Testing in Forest Tree Breeding, FAO-North Carolina State 
Forest Tree Improvement Training Centre. 

Van Buijtenen, J.P., 1976. Mating Designs, IUFRO Joint Meeting on Advanced Generation 
Breeding, Bordeaux. 

Participants visiting the CVG plantations 
of Pinus caribaea in Uverito 

- 160 - 


W.H. Barrett 
Fiplasto S.A., Buenos Aires, ARGENTINA 


Introduction 16 1 

Concept and design 161 

Definitions 161 

Historical background 161 

Planning and design 162 

Clones or seedlings 162 

Number of clones or seedlings 162 

Initial planting distance 163 

Experimental design 163 

Vegetative propagation 163 

Establishment 1 64 

Location 164 

Site preparation 1 64 

Planting 164 

Management and harvesting 165 

Cultivation techniques 1 65 

Plant cover 165 

Fertilization 165 

Pruning and thinning 166 

Protection . . . 1 66 

Flowering, pollination and handling of pollen 166 

Harvesting of fruit and seed 1 66 

Seed production 167 

Seed extraction and utilization ... 167 

Keeping records 167 

Bibliography 167 

\J The text of this article has followed the structure and content of the book, "Seed 
Orchards", edited By R. Faulkner (1975) for the IUFRO (International Union of Forestry 
Research Organizations) Working Group S 2-03-3, with additional information taken from 
lectures given by D.G. Nikles in Kenya in 1974 and A.G. Brown and K.G. Eldridge in 
Australia in 1977. 

- 161 - 


The modern trend of producing wood and wood products from plantation forests implies 
the availability of enormous quantities of seed, if extensive areas of forest are to be 

It is sometimes difficult to obtain enough high-quality seed in sufficient quanti- 
ties for the increasingly large-scale expansion plans in the different forest regions 
of the world. 

Moreover, the cost of forestry activities, from the preparation of the soil to the 
felling of the trees and extraction of the logs, and also the irrecuperable factor of the 
long wait until the trees are ready to harvest means the breeder must know in advance, 
that the seed will be good. 

Seed orchards ensure a supply which is regulated by the foresters themselves, and 
which provides steady, reliable production for their plantation programmes. 

The establishment of seed orchards is therefore accepted as a necessary forestry 
activity, which should be started as soon as possible in afforestation programmes relying 
on seed supplies. 


According to the general definition by Faulkner (1975), seed orchards are the 
most important means the tree breeder has for mass-producing seed for large improved 
plantations, based on the best selected trees. 

The classic definition by Zobel e_t al (1958) states that "A seed orchard is a 
plantation of genetically superior trees, isolated to reduce pollination from genetically 
inferior outside sources, and intensively managed to produce frequent, abundant, easily 
harvested seed crops. It is established by setting out clones (as graftings or cuttings) 
or seedling progeny of trees selected for desired characteristics". 

In some cases, seed orchards are established to mass-produce seed of some popula- 
tion of which it is impossible to obtain seed in adequate quantities, without too much 
attention being paid to the genetic superiority of the individuals. For this reason, 
the Organization for Economic Cooperation and Development broadens the concept of the 
seed orchard for the use of international trade, as follows (Brown and Eldridge, 1977): 
"A seed orchard is a plantation of selected clones or progenies which is isolated or 
managed to reduce pollination from outside sources, managed to produce frequent, 
abundant, easily harvested crops of seed." 


Although the concept of a seed orchard was applied before 1940 to other tree 
species such as rubber and quinine, it was only in 1949 that the first pine seed orchards 
were planted in Sweden. The use of this technique was intensified as from 1957 in the 
south-eastern United States, in line with the expansion of pine plantations in response 
to the establishment of many large paper factories in the region. The Cooperative 
Programme of the University of North Carolina alone in association with 32 forest in- 
dustry ventures in the region, has a total of 1 700 ha of pile seed orchards (185 
orchards in three States) . This scale is justified by the seed needs of an annual 
afforestation programme of 180 000 ha (Sprague ejt al , 1978). 

The United States Forest Service figures for 1972 (Wright, 1976) list about 3 000 
ha of Pinug taeda and . elliattii seed orchards that could produce seeds for at least 
300 000 ha each year. 

Howvr, the Queensland Department of Forestry in Australia can claim the honour 
of having had the first fully productive seed orchard. The dona 1 orchard of Pinus 
elliottii near Bttnrah has been producing at least 20 kg/ha/yr since 966. Thi. is the 
**d used fey the D*parttnt for its plantations (Brown and Eldridga, 1977). 

- 162 - 

Another species that has been used intensively in afforestation programmes in 
Australia and New Zealand is Pinus radiata. These countries have 625 ha of clonal or- 
chards from grafts, producing more than 4000kg of seed per annum (Brown & Eldridge, (1977) 

In 1975 second generation seed orchards began to be planted in the United States 
and in Australia. 

To date, almost all the information available concerns conifers, especially pines. 
Forest hardwood seed orchards are recent and on a smaller scale; Tectona has been 
grown in New Guinea, Thailand, Nigeria and India, and Gmelina in Nigeria. Mention should, 
however, be made of work on eucalypts in Australia, South Africa, Morocco, Portugal and 
more recently, Brazil, where private enterprises, with the collaboration and advice of 
IPEF (Institute de Pesquisas e Estudios Florestais), are cooperating on a major effort 
in this field. 

World experience (despite the fact that this technique is relatively new) and the 
results obtained with different species in different regions indicate enormous potential 
for seed orchards for better yields of forest crops, and, therefore, more effective use 
of sites and cheaper forest production. 


Generally speaking, the design of a seed orchard is based on the assumption that 
all the clones or progeny composing the orchard will flower at the same time, be com- 
pletely inter-fertile with all their neighbors, produce the same amount of pollen and 
ovules, yield the same amount of seed, have the same degree of compatibility when 
grafted, the same type of growth and a similar form of crown, etc. 

Experience shows that this is seldom the case, since for each species and each 
region, these characteristics must be studied in detail and all possible information 
collected on clonal behaviour, compatibility, combining ability and any other information 
inherent in the production of seed and its future behaviour, in order that the design 
of the next and subsequent orchards can make maximum use of the available data. 

Clones or seedlings 

There is a wealth of literature on whether to use clones or seedlings from selected 
trees, (Toda, 1964)., but in general it is agreed that only in special circumstances is 
it desirable to use seedling progenies. Unless control pollinated seed is available, the 
offspring will very likely be genetically inferior to the parents. Secondly, selections 
in the orchard are applicable only to this environment. N ikies (1974) agrees that in 
certain cases the seedling seed orchard has a special place in improvement programmes 
when working with species that root with difficulty and that may flower and set fruit 
early. The same author maintains it is a mistake to believe that the seedling seed 
orchard is the low-cost solution for small breeding programmes, since great expertise 
is essential to a successful orchard. 

Many aspects of the establishment, design and management of seed orchards are 
common to both clonal and seedling seed orchards. However, there has been more ex- 
perience with the former, basically with grafting, although rooted cuttings have also 
been used. 

In practice, very few seedling seed orchards from selected trees have been estab- 
lished. The decision is made not on genetic grounds but springs rather from a reluctance 
to use a little-known technique where the investment is bound to be long-term. 

Number of clones or seedlings 

It is generally agreed that the right number of clones or families to obtain the 
maximum genetic gain and avoid the bad effects of inbreeding, is between 15 and 20. There 
are, however, good reasons for establishing a greater number of pheno-types to start 
with. Possible incompatibility between the graft and the stock, flowering habits, and 
also the need to thin clones, (leaving those with better general combining ability), 
justify starting a seed orchard with a greater nunber of clones than is usual (in general 
practice 60 to 100 clonesjl "~ 

- 163 - 

According to Lindgren (1974), genetic gain is maximized by using a smaller number 
of clones , provided they flower simultaneously and their orchard output is not used for 
further selection. Five clones provide enough genetic variation to ensure adaptation to 
different sites, resistance to diseases, changing uses for wood, and so forth. 

An extreme illustration of this concept is the use of two clone seed orchards with 
high specific combining ability for maximum genetic gain of known geno-types. 

On the other hand, for progenie of selected trees, Shelbourne (1969) has advocated 
starting the orchard with more than 200 families, with intensive intra- and inter-family 

Initial planting distance 

On the assumption that the best final spacing for most cultivated species is 10 
metres, depending on expected intensity of selection, or earlier fruit-setting through 
better pollination, the initial distance can vary from 2 to 6 metres. This spacing has 
been used successfully for pines. As a general rule, it may be said that the space 
between plants should be sufficient to allow free and complete crown development, with 
maximum sunshine, to ensure a plentiful crop of fruit for several years before thinning 
becomes necessary. 

The minimum distance is often limited by the space needed for mechanical control 
of weeds and grass, applying herbicides and fungicides, and seed collection. 

Planting can be in squares, rectangles or quincunxes; Nikles (1974) mentions a 
hexagonal distribution for better mechanical seed collection. However, after selection 
and culling of undesirable genotypes, the final distribution is completely irregular. 

Experimental design 

The design most widely used throughout the world for clonal orchards and seedling 

seed orchards is the randomised complete block. Shape and distribution within the 4 block 

should be such as to ensure cross pollination in which all clones participate. To avoid 

self ing, it is desirable to alter the random design where ramets of the same clone are 
close together. 

Some systematic distributions can be recommended within the block when the initial 
number of clones is maintained until the end, otherwise the thinning of clones disrupts 
distribution as much as random design. 

The size of the orchard depends on the forester's need for seed; in other words, 
it is in direct proportion to the projected afforestation annual programme. 

The number of blocks will depend on the quantity of seed that experience of the 
species and site dictate. 


As indicated above, most orchards are established by using clone ramets. The tech- 
nique commonly used is grafting, with various methods depending on the species and the 

Some species, mainly due to incompatibility between stock and graft, cannot be 
successfully propagated by this method, and therefore other techniques must be used. 
This was the case with some widely cultivated species such as Pinus radiata, Eucalyptus 

S rand is and Pseudotsuga roenziesii, in different regions of the world. Sometimes, the 
ecline and death of the clone take place over a fairly long period of time, complica- 
ting the operation of the orchard and post-poning definitions. With Pinus radiata, 
the problem was solved by making use of the ease with which its cuttings take root, and 
the same method was followed with Eucalyptus grand! s, although seedlings from selected 
trees have also been used to establish seed orchards. Considerable efforts have been 
made with Douglas fir to improve grafting techniques, detect incipient incompatibility, 
replace defective grafts and find compatible roots tocks. With this species, seed 
orchards were often established from seedlings. 

- 164 - 

The methodology, techniques and problems of vegetative propagation have already 
been discussed in another lecture. 


In deciding on the location of a seed orchard, the following factors should be 
borne in mind: climate and soil for high seed production; isolation from undesirable 
pollen sources; facility of access and nearness to work areas. 

The ecology of the site where the seed orchard of a particular species, variety 
or provenance is planted, should be such as to encourage optimum development and maximum 
fruit setting. Where the species are native and the orchard is located within the 
range of the species, the best site can probably be identified quite easily. With the 
more extensively cultivated species, there is already sufficient knowledge available to 
reduce the danger of choosing a bad site. However, for species or provenances of recent 
introduction and little history of. cultivation in other regions, choice of the seed 
orchard site should have top priority. 

As regards climate, it is usually unwise to select sites exposed to extreme tem- 
peratures, or low areas exposed to frost. Some species, like Pinus elliottii and 
tae< * a > d not appear to be affected by slight changes, whereas . caribaea is very 
sensitive to frost. 

Although one may generalize with regard to the best texture, quality and fertility 
of soils for better fruit production, each species may have different requirements in 
this respect. 

Concerning the isolation factor, it is difficult to achieve perfect isolation 
(which would be theoretically acceptable) and completely avoid contamination by un- 
desirable pollen. The orchard would have to be situated outside the forest area or in 
a remote or not easily accessible place. Experience shows that, in orchards with normal 
flowering and fruiting, complete isolation can be sacrificed in favour of better siting 
as regards accessibility and nearness to working places. The minimum distance accept- 
able varies with species, topography, winds, etc. It has been found that for Pinus 
caribaea var. hondurensis a distance of 200 metres is sufficient (Nikles 1974) to 
achieve adequate isolation . 

The site should be protected from wind, fire and animals, and near places of easy 
access and intensive work such as nurseries. The site must be visited after for manage- 
ment, fertilization, irrigation, pruning, collection of branches for propagation, polli- 
nation, and regular inspection. Pest and disease control are also easier. Proximity to 
labour force also facilitates seed collection, drying and extraction. 

Site preparation 

The site selected for the orchard should be prepared according to normal affores- 
tation practices, with no skimping on preventive control of rodents, ants and other pests, 
Clearing and weeding should receive the same care lavished on intensive agricultural 
crops . 

Site preparation should facilitate mechanical work as much as possible. Paths, 
fire-breaks, etc., should be methodically laid out. 

In clay or poorly-drained soils, it is desirable to chiselp lough the planting rows. 

It is also desirable to use land with little or no slope; however, if no flat ground is 

available, contour ploughing, or, in extreme cases, terraces, to check or prevent soil 
erosion is recommended. 


The primary objective should be maximum survival of the material planted. If this 
material consists of nursery plants (seedlings of selected Plants), or grafts in pots 

- 165 - 

or clumps of soil, planting should be timed to coincide with optimum weather and soil 
moisture conditions to ensure maximum success, as replanting in subsequent years has 
proved to be ineffective. It is therefore desirable to plant a greater number of ramets 
closer together. 

On some occasions, good results have been obtained by planting the future root 
stocks straightaway in the spot where they will be growing permanently and later grafting 
III jiitu. This is particularly true for species that do not take well on transplanting. 
In this case it is advisable to plant 2 or 3 roots tocks in each spot, as a precaution 
against unsuccessful grafts or incompatibility, or in order to select the best graft, etc. 
This latter method has produced good results in Queensland, with 85 to 100% success in 
grafts of Pinus caribaea and Aj^aucaria cunninghamii , and also in the southeastern United 
States with Pinus elliottii and . taeda. It has the advantage of ensuring more rapid 
growth, earlier flowering and no root curling. (Nikles 1974). Among the disadvantages 
are the need for more experienced grafters and more frequent visits to tend the grafts 
in the field; also, weather conditions cannot be controlled (temperature, winds). 

When dealing with little-known species, unpredictable climates or unskilled staff, 
it is therefore preferable to use material which has been grafted, or has already taken 
root, in shelters or nurseries. 


Cu 1 1 i vat i on t e chn ique s 
Plant cover 

It is advisable to weed the plantation the first year, especially around the shoots 
or seedlings, but experience has shown that later on it is better to let the grass grow 
naturally (provided there are no aggressive rhizomes) or to sow grass, which kept 
short. This practice reduces or prevents soil erosion; affords excellent fire protection, 
adds organic matter to the soil and keeps the orchard in excellent condition for work. 

Weed killers or a mulch of straw and/or manure around the plants following super- 
ficial raking - never deep raking or hoeing that could damage the roots - are recommended 
when drought makes weeds over-competitive. 

In areas with periodic droughts, or in dry years, it is preferable to irrigate the 
plantation to conserve soil moisture rather than leave the ground bare (Nikles 1974). 


Fertilization can be very effective in increasing seed production on skeleton or 
poor soils. The type, rates and timing of fertilizer applications, vary according to 
species and site. Experiments described by Nikles (1974) are summarized below. 

In the region of origin of Pinus elliottii, the equivalent of 1 kg of nitrogen 
(3 kg of ammonium nitrate) per tree is applied annually at the end of the spring or the 
beginning of the summer. On phosphorus deficient soils, (i.e., less than 8 to 10 kg 
of extractable P 2 05) and potassium (less than 60 kg of K 2 per hectare), 2,5 kg of 
super phosphate and 1 to 1.5 kg/tree/yr of -potassium chlorate are applied. Lime is also 
used where acidity is high or there is a deficiency of Ca and Mg. 

In Queensland, an appreciable increase in seed production has been achieved through 
NPK applications. 

For Pi nu * taeda, the Cooperative Programme of the state University of North Carolina 
has reached the conclusion that the best way to achieve good seed production is to ferti- 
lize (NPK) in years of normal rainfall, and irrigate during periods of drought. 

In Texas specific recommendations for fertilizer applications are made in the 
light of results of leaf analyses. It is a common practice to apply nitrogen to promote 

- 166 - 

Pruning and thinning 

Neither pruning nor crown reduction nor girdling have produced the anticipated in- 
crease in seed production of species with high apical dominance such as Pinus elliottii, 
ZJL tae< * a an< * radiata. On the other hand, bending down TP. radiata branches has proved 
successful in facilitating cone collection; as has pollarding which forces the tree to 
produce cones lower down. Low branches are only pruned to facilitate access and move- 
ment within the orchard. 

Clonal selection in the orchard will depend fundamentally on the result of the 
progeny tests, which will eliminate clones with poor general combining ability. The seed 
production of each clone should also be borne in mind. If the clone produces not seeds 
but male flowers (pines) and has good combining ability, it should not be culled. In 
the selection of clones, the degree of incompatibility and other factors producing weak 
or low-yield ramets should also be borne in mind. 


Seed orchards are highly specialized, expensive and very important for afforesta- 
tion programmes, so they are far more carefully protected than ordinary forest plantations, 
This is not difficult, since the orchard is frequently visited by trained staff who can 
easily spot any problem that could affect the normal development of the orchard. Damage 
from diseases, insects, animals and birds, machine damage to trunks and branches and 
wind or frost damage can thus be remedied or prevented. 

Fires should be prevented or controlled, by keeping fire-breaks clear, not allow- 
ing weeds to grow too tall, and having fire- fight ing teams always ready for emergencies. 

Measures should also be taken to avoid erosion and excessive compaction of the soil. 
Flowering, pollination and handling of pollen 

The seed orchard manager needs a good understanding of the environmental factors 
controlling flowering and seed set, the movement of pollen and the degree of inbreeding 
and outcrossing. These are aspects that need to be studied in greater detail, and the 
reader is therefore referred to Chapters 7, 8, 9 and 11 of Faulkner's book (1975). 

Until now, pollination has seldom been considered sufficiently reliable for proper 
seed setting to reduce self ing and avoid undesirable pollen. Brown and Eldridge (1977), 
hold that it is possible to pollinate 1 000 flowers with 6 of pollen. One man with simple 
equipment can pollinate 5 000 flowers daily. They also maintain that this technique, 
with some refinements, will soon become part of seed orchard management, particularly 
when the orchards are young. 

Harvesting of fruit and seed 

Harvesting techniques depend on species, climate, site, size of the orchard, local 
or national economic conditions of seed. Mechanized harvesting is justified if labour 
is scarce or is expensive, or if very fast harvesting is necessary because the seeds 
ripen and are dispersed quickly in extensive orchards. 

Some pines species have cones that fall very easily. These can be harvested by 
hand, or with mechanical tree shakers. However, in some serotinous species the cones 
are not easily detached from the tree. In these cases, elevating work platforms can 
be used. The Australian experience with P. radiata indicates that 6 kg per day can be 
harvested with this machine whereas only T.8 kg of seed/man/day can be collected by 
hand (K. Willcocks in Brown and Eldride 1977). In the southeastern United States, 
studies are being made on the use of a suction harvester to pick up the seeds of Pinus 
taeda from the ground. Where harvesting time is limited because of rapid dehiscence 
of the cones or fruits, they can be harvested green so long as the orchard staff are 
familiar with off-tree fruit ripening techniques. 

With eucalypts, that coppice well and produce a lot of seed, Brown and Eldridge 
(1977) recommend felling the tree to facilitate harvesting. With serotinous cone 
species that have very little fruit, and where labour is expensive, harvesting is at 
two or three-year intervals to reduce costs. 

- 167 - 

Seed production 

This varies with the species, age of the trees, clonal composition, site, management, 
etc* It i* important for any organization managing an orchard to determine production 
potential, so as to match orchard size to the forestry programme. Nikles (1974) reports 
that a high-yield seed orchard of Pinus elliottii in Queensland produced 145 Ibs/acre 
(approximately 160 kg/ha); a normal yield for this species in Australia is 75 Ibs/acre 
(83 kg/ha). With Pinus caribaea var. hondurensis the experience of the Byfield orchard 
indicates that between 13 and 34 Ibs/acre/yr can be obtained, with an average over 8 years 
of 23 Ibs (25 kg/ha). Other younger orchards obtained lower seed productions. The 
average yield of Pinus radiata, after 10 years, is 20 Ibs/acre (22 kg/ha). In the United 
States, the average for Pinus^taeda, a higher yielding species than . elliottii is 
estimated at 55 kg/ha/yr. 

Seeds from orchards are usually bigger, and have better germination and viability 
than seeds from natural forests or plantations. 

Seed extraction and utilisation 

When cones are collected from an expensive seed orchard, maximum care should be taken 
in extraction and subsequent operations so as to maximize viability. The seed should be 
extracted at low temperature under dry conditions. Seed should be stored in accordance 
with the requirements of each species, with adequate protection from insects and rodents, 
and proper disinfection. 

Keeping records 

Efficient seed orchard management and utilisation means keeping careful records, 
including: charts of ramet location, listing cultivation techniques, fertilization, costs, 
phenological observations, incompatibility, fruit production, relative seed production 
from each clone, etc. Experience has shown that the more information recorded, the better 
the management and future upkeep of the orchard. 


Brown, A.G. & K.G. Eldridge, 1977. Seed orchard design and management. International 
Training Course in Forest Tree Breeding, Selected Reference Papers, Australian Develop- 
ment Assistance Agency. Canberra, pp. 147-154. 

Faulkner, R. 1975. Seed orchards. U.K. For. Comm. Bull. N. 54. London. 

Lindgren, D. 1974. Aspects of suitable number of clones in a seed orchard. Proc. Joint 
IUFRO Meeting 52.04-1 ,-2, -3, Stockholm. 

Nikles, D.G. 1974. Seed orchards-concept, design, establishment and management. FAO/ 
DANIDA Training Course on Forest Tree Improvement, Kenya, pp. 176-195. 

Shelbourne, C.J.A., 1969. Tree breeding methods. Tech. Paper no. 55, For. Res. Inst. 
Rotorua, New Zealand. 

Sprague, J., J.B. Jett and B. Zobel. 1979. The management of Southern pine seed orchards 
to increase seed production. Proceedings of A Symposium on Flowering and seed develop- 
ment in trees. Ed. Frank Bonner, US Forest Service, IUFRO, Mississipi State University, 
Mayo 1978, pp. 145-192. 

TodaR. 1964. A brief review and conclusions of the discussion on seed orchards. Silvae 
Genetica 13 (1-2): 1-3. 

Wright, J.W., 1976. Introduction to Forest Genetics. Academic Press Inc. N.Y. pp.463. 

Zoble, B., J. Barber, C.L. Brown and T.O. Perry, 1958. Seed orchards, their concept and 
management. J. For. 56: 815-825. 

- 168 - 

M. Quijada R. 

Institute de Silvicultura 
Universidad de Los Andes 
Merida, Venezuela 



Definition and importance 168 

Crossing value . 1 69 

Estimation of genetic statistical parameters 169 

Evaluation of groups or lines 169 

Evaluation of individual trees 169 

Types of progeny 1 69 

Crossing systems 1 70 

Experimental designs 1 70 

Cultivation techniques 1 70 

Evaluation 171 

Records 171 

Basic select bibliography 171 

Annex I Specific models 1 72 

1. Variance analysis for a factorial genetic design in a completely random and equal 
replication statistical model. 

2. Variance analysis for a modified diallel genetic design and a completely random 
statistical model. 

3. Variance analysis for a genetic design of polymixia or free pollination in a 
completely random statistical model. 


The purpose of progeny tests is to estimate: (I) The genetic value of an individual 
on the basis of the behaviour of its offspring, or (2) the genetic value of individual 
half-siblings or full-siblings. 

In programmes based on phenotypic selection, where environmental influence is un- 
known and where selection is consequently not very reliable, progeny tests are indispens- 

- 169 - 

Progeny tests facilitate: 

(a) Evaluation of the crossing value (combining ability) of trees; 

(b) Estimation of genetic statistical parameters (variance, correlation, etc.). 

(c) Evaluation of sub-populations or lines for specific uses in plantations; 

(d) Evaluation of individuals for the purpose of continuous selection. 

Crossing value 

This is the value of a tree, determined by the average of its progeny obtained from 
crosses with one or several other trees, following a specific plan. The crossing value of 
individuals is determined by specific crossing systems (genetic designs). From this, we 
can determine the combining value of an individual which is a statistical parameter that 
indicates the ability of an individual to transmit certain of its characteristics to 
progeny resulting from crosses with one or more other individuals. This information is 
used as a basis for selective thinnings in seed orchards. 

Estimation of genetic statistical parameters 

This is done through a series of theoretical assumptions on the components of the 
phenotype and constitutes the mathematical basis of genetics. It involves the use of 
special types of genetic and experimental designs. 

These require determination of the components of variance and co-variance, using 
Expected Mean Squares in Variance and Co-variants Analysis Tables. 

The determination of such parameters as heritability (h2) , makes possible eventual 
estimation of genetic gains or the response to breeding schemes as regards both how many 
in numerical value and how much improved. 

Evaluation of groups or lines 

In specific cases, the results of particular combinations between clones or families 
resulting in progeny outstanding for a specific characteristic (high value of specific 
combining) can be used. 

Evaluation of individual trees 

For the purposes of continuous selection, individual evaluation is made as and when 
advanced seed orchards are established, i.e. in future generations. 

This requires continuous selection within progenies of the most outstanding individ- 
uals. It can lead to problems of inbreeding in orchards of an older generation, but 
this can be counteracted by having a sufficiently broad basis for initial selection, and 
separating the seed production population from the breeding population (see paper entitled 
"Planning and Strategies of a Forest Tree Improvement Programme 11 ). 


The quantity of information that can be provided by trials depends on the type of 
progeny, among other things. 

We may distinguish two principal types: semi-fraternal (half-sibs), and fraternal 
(full sibs). 

Half-sibs are progeny with a known parent in common (the mother tree). They are 
typically freely pollinating, the exact degree of inbreeding is unknown. 

Information obtained will depend on the value of the seed tree. 
I/ For more detailed information, see paper entitled "Quantitative Genetics", 

- 170 - 

Full sibs are progeny with two known parents in common. This can only be achieved 
through controlled pollination. 

In this case, the tests will produce more information, as it is based on the values 
of both parents. However, most progeny tests made to date have been with half-sibs, since 
the establishment of full-sib tests involves considerable work and expenditure. 


The way in which parent trees are combined influences the results of tests, since 
it determines the type of progeny obtained. 

There are two important general types of crosses in the field of forestry: (1) 
crossing systems with an unknown father, and (2) crossing systems with a known father. 

In the first case the principal systems are: (a) free pollination and (b) poly- 
crossing; in the second (c) diallel designs (complete, modified, partial diallels), 
(d) factorial designs, and (e) mating of a single pair. For detailed information on 
these designs, see the paper entitled "Systems and Designs of Controlled Crossing". 


The environmental component in progeny tests is controlled by the use of experi- 
mental designs which seek to reduce the non-genetic effects. Randomization (random distri- 
bution), replication (repetition) and local control (blocks) should be taken into con- 

The use of blocks facilitates control of local variability, exposing the progeny 
to fairly standard site conditions. Ideally, replication makes it possible to expose 
individual progeny to different site conditions, which occur quite frequently in -the same 
plantation. The genetic and environmental components of the phenotype can thus be 
estimated. Randomization is indispensable for obtaining estimates not vitiated by ex- 
perimental error. 

The testing site should be representative of the area where the seed will eventually 
be used in a normal plantation. The best experimental design usually depends on the 
number of progeny, plot size, and site variability. The complete blocks design is often 
used for low numbers of progeny and trees per progeny, and sites of average to low varia- 
bility. For a large number of progeny (more than 25), large plots and very heterogeneous 
sites, it is advisable to use incomplete block designs such as lattices. 

Plot size can range from one tree per plot upwards. Low numbers, particularly in 
row plots, are desirable if the value of intra-progeny competition is to be ascertained. 
High numbers are desirable with square or rectangular plots in estimating genetic statis- 
tical parameters where the value of competition is a bias factor. In this case, one con- 
dition is a variable number of trees in the centre of each plot of a given progeny, that 
are not in direct contact with other progeny, or else are surrounded by buffer areas of 
trees of their own kind. 

In order to cover the maximum intra-progeny variability, it is recommended that the 
number of replications be increased for smaller plots. 

Replication can be spatial (control of local environmental variability) and temporal 
(control of climatic or biotic variations). 


Spacing should be the usual planting distance in a normal programme. Tending 
should, as far as possible, be the same as for large-scale plantations. "Luxury tending" 
should be avoided. 

Protection against fire, wild animals, etc., should be effective. 
\J For more detailed information on experimental designs, see the paper on this subject. 

- 171 - 


The characters evaluated will depend on the actual and potential uses of the species 
under consideration. There are, however, several (general) characters common to the 
majority of uses, such as survival, growth and vulnerability to pests and diseases; other 
(specific) characters will depend on the use envisaged. 

Evaluation in the field usually begins six months or a year after the test, with 
mortality and attack counts, and measurement of initial growth. Evaluation is usually 
repeated in the second and in the fifth, seventh and tenth years, including other charac- 
teristics in the measurement, as the individual trees develop. For most characters, the 
minimum age for making reliable evaluations is considered as equal to one-third of the 
final rotation of the species. For some characteristics, such as wood quality, it is 
usually necessary to wait until the end of the rotation, with periodic evaluations every 
three or five years. 

Evaluations are also made in the nursery, for correlative comparisons with field 
behaviour. These evaluations can also be used for selecting within the nursery itself. 
Even more important, these studies can include seed characteristics such as weight and 


A detailed record of the material and the procedures used is indispensable, a file 
should be kept for every test made. This file should include a summary of the objectives 
of the test, and all observations made during the course of the test, data and analysis 
of evaluations, etc. This procedure will make it easier to come to the right conclusions 
at the end of the test. 


Becker, Walter A., 1967. Manual of Procedures in quantitative genetics. Washington State 
University Press, Pullman, Wash., USA. 130 pp. 

Cockerhan, C. Clark, 1963. Estimation of Genetics variances. Statistical Genetics and 
Plant Breeding. National Academy of Sciences, Publication 982, USA. pp. 53-94. 

Falconer, D.S., J960. Introduction to Quantitative Genetics, Ronald Press, 1971 New York, 
365 pp. 

Franklin, E.C., 1969. Quantitative inheritance, Heritability and Combining Ability. 

Forest Tree Improvement Training Course, FAO-N.C. State University, Raleigh, N.C. 
USA. Volume I: 96-107. 

Mather, K. and Jinks, J.G. 1971, Biometrical Genetics Chapman & Hall, N.Y. 

Matheson, A.C., 1977. Progeny Tests. Training Course in Forest Tree Improvement, 
Australian Development Assistance Agency, Canberra, Australia; pp. 113-119. 

Namkoong, G., E.B. Snyder, and R.W. Stonecypher, 1966. Heritability and gain concepts 

for evaluating breeding systems such as seedling orchards, Silva Genet ica 15: 76-84. 

Roberds, J.H., 1969.' Progeny Testing in Forest Tree breeding. Forest Tree Improvement 
Training Course, FAO-N.C. State University, Raleigh, N.C. USA Volume I, 123-135. 

Wright, J,W., 1976. Introduction to Forest Genetics. Academic Press, Incl. New York. 
463 pp. 

- 172 - 

Specific models 

Variance analysis for a factorial emetic design in a 
completely random and equal replication statistical model 

Source of 

Degrees of 
freedom df 

Mean square 
expectation E(MS) 



m - 1 

<T e +rmh+rhm 

o* m - Cov(fm) 


h - 1 

(T e + r mh +rmh 

6 h 2 - Cov(fh)- 1 VA 

Male x Female 

(m-1) (h-1) 

<T e + r m h 

CT m hVarCf)-Covfm 


mh (r-1) 

*e 2 

- Cov f h 


Total mhr- 1 





Component of variance for the source of variation indicated by the letter (s) 

Co-variance of half-sibs with a male parent in common. 

Co-variance of half-sibs with a female parent in common. 

Sibling variance. 

Additive component of genotype. 

Non-additive component of genotype. 

- 173 - 


Variance analysis for a modified diallel genetic design 
in a completely random statistical model 

Source of 

Degrees of 

Mean square 


General combining 

p - 1 

ae 2 + r e 2 + r (p-2) g 2 

ag -Cov f 

Specific combining 

Experimental error 

p(p-3) 12 \ 0e 2 + r e 2 
p(p-l) (r-l)/2 oe 2 

ere 2 Var f-2Cov f 



Cov f- 
Var f- 



Number of parents. 

Components of variance for the source of variation as indicated. 

Half-sib co-variance. 

Full-sib variance. 

Additive component of genotype. 

Genotype component. 

Analysis of variants for a genetic design of polymixia or 
free pollination in a completely random statistical model 

Source of 

Degrees of 

Mean square 



(P - 1) 

2 2 
cre z + r p 

a p Cov fl 

4 VA 

Experimental error 

f(r - 1) 

oe 2 

* Component of variance for the source of variance indicated, 
Cov f Co-variance of half-sibs. 
VA* Additive component of genotype. 

- J74 - 

M. Quijada R. 

Institute de Silvicultura 
Universidad de Los Andes 
Merida, Venezuela 


Existence and importance 

Rank position 

Variance analysis . . . 

Regression analysis 




Genotype-environment interaction may be defined as lack of uniformity in the res- 
ponse of two or more groups of plants grown in two or more environments; one group may 
show the best growth in one environment, but be mediocre in another. 

The interaction may show itself in a change in the relative position of the groups 
in each environment , or in differences in the degree of superiority, even when the rank 
positions are equal in each environment. 

In graph form, an interaction is represented by lines which cross or approach each 
other, as the environment varies, following a given gradient. 

Value of character tested 

Value of character tested 

1 2 


1 2 


Since it is more apparent, the first makes it necessary to select a distinct group 
for each environment, in order to obtain the best gain in productivity. 

The second is important as a guide for the extrapolation of data to environments 
different from those tested. 

The presence of genotype-environment interaction reflects the existence of environ- 
mental variations, even over relatively small distances, and of variations in the require- 
ments of different genotypes in both macro and micro environments. The same reasoning 
may be applied to seasonal variations. 

- 175 - 

This is the fundamental reason why experiments need to be repeated both in different 
areas (control of local environmental variation) and at different seasons (control of 
cyclical variation). 

Different levels of genotype-environment interaction may be distinguished in forestry, 
due mainly to the genotypic constitution in question. Thus we note the existence of 
species x environment, provenance x environment snf progeny x environment effects. 

In terms of species, the differences with regard to adaptability to environments 
are greater, i.e. there are species which possess a wide range of adaptation, so that the 
detection of interaction necessitates tests over a wider range of environmental conditions. 
If the genetic spectrum, i.e. the provenances and progenies, is reduced, interaction tests 
become more sensitive, in that the ranges of tolerance narrow, and this sensitivity shows 
up in different reactions to minor variations in the environment. 

Of the environmental factors that indicate interactions, the most tangible are soil 
conditions, since these can be detected within a plot, over relatively small areas. 
Climatic conditions have to be more extreme in order to be detected, e.g. variations due 
to extreme conditions of temperature (low or high), humidity (drought or flooding), etc. 

The interaction may be considered as an indicator of the relative stability of a 
genotype. If the interaction is close to zero, the genotypes are sufficiently stable for 
the characteristics under consideration, i.e. their relative positions and the differences 
in magnitude of responses are similar in the different environments; the only thing that 
changes is the actual magnitude of the responses, which will depend on the particular 
genotypes (species, provenances, individuals) and sites under consideration. 

Genotypes with a wide range of adaptability, as shown by zero or insignificant 
interaction, are called plastic, while those with a high susceptibility to marked varia- 
tion in behaviour with slight variations in the environment are called rigid or strict. 

Since the components of genotype and environment also play a part, though small, 
in the response of individuals (evaluated in the pheno types), the interaction will never 
represent 100% of the variation observed; however, the nearer one comes to this value, 
the lower will be the stability of the genotype in different environments, and from the 
point of view of the breeder, the less reliable it will be for large-scale propagation 
under normal plantation conditions. It will be more suitable for plantations in limiting 
environmental conditions, since these are established in relatively small areas and under 
more honogeneous site conditions. 

The present tendency of breeders is to produce sub-populations with a broad spect- 
rum of adaptability, which, in the long run, can represent greater gains for lower costs. 

If the interaction is not significant, selection will be made on the basis of the 
mean of the responses of the genotypes in all the environments. This same criterion may 
be adopted when the interaction is significant, but at levels very close to the lowest 
critical limit of probability (usually 5%), in which case it may be ignored in practice. 

The practical problems of using interaction as a yardstick are highlighted by 
certain factors, the difficulty being to interpret it correctly. Very often the existence 
existence of interactions is reported on the basis of a very few tests, located very 
close to each other, which in practice would be of very dubious applicability from the 
point of view of costs. In these cases it is recommended that the interaction be con- 
sidered as part of the experimental error and the average yields from the sites tested be 
taken as the basis for work* 

Effective practical use of interaction requires essentially testing of the same 
genotypes in a range of environmental conditions which represent marked variations, with 
areas sufficiently large to be used for plantations. 

- 176 - 

Rank position 

A simple, preliminary way of evaluating the presence of genotype-environment inter- 
action is by using the rank position of each of various genotypes in each of various 
sites or environments. 

The genotypes are graded from best to worst, according to their yield in each site, 
assigning them a number from 1 to n (according to the number of genotypes). 

If there are noticeable changes in position from one site to another, this indicates 


Genotypes I 



A 1 
B 5 
C 4 
D 3 
E 2 





A disadvantage of this method is that rank classification gives us only n overall 
idea of the magnitude of the changes from one environment to another, since the changes 
will always be in constant units. In addition, there is no statistical information avail- 
able on the significance of these changes. 

A statistical approximation can be obtained by using rank correlation, using pairs 
of stations. These may be all the pairs possible according to the number of environments, 
or the environmental extremes in question. 

Variance analysis 

The importance and magnitude of the interaction are evaluated by using different 
methods of variance analysis. One method is to combine variance analysis with a specific 
experimental design (e.g. random complete blocks). 

Sources of variation 







Genotypes x 

(t-l)(L-l) MSI MSI 

rfe+Kl 61 

est. MS 

Errors L(t-l)(r-l) est .MS 

Total Lrt-1 6F 2 

The significance test gives us information on the overall importance of the inter- 
action. We obtain its magnitude (0I 2 ) and effective participation in the total variation 
(6F) by estimating the variance components, using the means squares ( E(MS) ) anticipated 
from the variance analysis table. 

- 177 - 

Regression analysis 

Another way of using the principle of variance analysis is through the regression 
of the genetic values in the environments. For this it is necessary to establish the 
same genetic entities (normally progeny) in various well-differentiated environments 
(preferably more than two) . The individual values of a given characteristic of each 
genetic entity (genetic values) are related to the overall means of each site (environ- 
mental values) adjusting simple lineal regressions. This results in various straight 
lines (as many as there are progeny), which are then compared with each other, analytically 
or graphically. The graphic form is the most usual, and the presence or absence of inter- 
action is indicated by the arrangement of the lines. If they are more or less parallel, 
the genetic entities are said to be relatively stable (no interaction). If the lines 
cross or markedly diverge from the parallel position, this indicates the presence of 

Analytical methods involve comparison of the lines to see whether they are to be 
considered as having the same slope, and involve, for each genetic entity quantifying the 
importance of the regression or deviation. 

The most stable genetic entities are those whose regression approaches the unit and 
the deviations zero. If these stabilities are associated with acceptable productivity 
(best or nearly best), then they will have preference over others which, although the best 
in a given environment, are unstable as a whole. 

Allard, R.W., 1964. Principles of Plant Breeding, John Wiley & Sons, Inc. New York. 485 pp. 

Bur ley, J. & Wood, P.J. (Eds), 1979. Manual on species and provenance research, with 

special reference to the tropics. CFI Tropical Forestry Papers No. 10 and 10A, Oxford, 
pp. 127-130. 

Falconer, D.S., 1960. Introduction to Quantitative Genetics. The Ronald Press Company, 
New York. 365pp. 

Goddard, R.E., 1977. GE interaction in slash pine. Third World Consultation on Forest 
Tree Breeding, Canberra, Australia, pp. 761-772. 

Hill, J., 1979. Genotype-environment Interaction - A Challenge for Plant Breeding. J. 
Agri. Sci. Ji5: 477-493. 

Namkoong, G., 1978. Choosing strategies for the future. Unasylva 30. (1 19/120) , pp. 38-41 . 

Rink, George. Variation in open-pollinated progeny plantations of Virginia Pine (Pinus 
virginiana Mill) . PhD Thesis, University of Tennessee, Knoxville, Tenn. U.S.A. 143 PP 

Participants visiting CVG Pinus caribaea plantations in Uverito 

- 178 - 


C. Palmbcrg 

Forest Resources Division 

Forestry Department 




Variability of the pathogen 

Variability of the host 

Principles of selection and breeding for disease resistance 

Assessing the control alternatives 

Developing methods for evaluation of resistance 181 

Selecting for disease resistance 181 

Breeding for disease resistance 1 82 

Maintaining resistance 183 

Practical implications . 183 

References 1 85 


A disease involves harmful physiological changes in a plant. The changes may be 
caused by non-pathogenic agents such as adverse climatic or soil conditions, or by a 
pathogen (Tarr 1972). The most important biotic diseases of forest trees are caused by 
fungi (Cowling 1969). Research on the genetics of insect resistance and on resistance to 
bacteria and viruses has not been nearly so extensive as on the genetics of resistance to 
fungal diseases. Nevertheless, enough is known to indicate that the inheritance of 
resistance to these agents differs in no major way from inheritance of resistance to fungi 
(Allard 1964; Painter 1966). Major stress in this paper will be placed on infectious 
diseases caused by fungi. The term 'breeding* will be treated in the broad sense, thus 
including procedures of selection. 

The general objective of a program to produce pest resistant trees is to reduce 
losses to a pest by genetic manipulation of the host trees (Callaham e ail. 1966). The 
probabilities of success in a breeding program will, to a large degree, depend on the 
establishment from the outset of a clearly defined set of objectives and a realistic 
order of priorities. The prospects for significant improvement of disease resistance 
are better in forest trees than in most cultivated crops, because forest tree species 
generally represent natural cross-pollinated populations which provide a wide range of 
genotypes for basic selection (Schreiner 1966). Concurrently with improvement in disease 
resistance, it is necessary to provide for improvement in yield and quality of the crop, 
A realistic weight must be given to the importance of each factor in the breeding program 
(Borlaug 1966). 

Selection and breeding for disease resistance does not vary fundamentally from 
breeding for any other character (Allard 1964; Dyson 1974). Consequently, any of the 
various methods of breeding appropriate for the crop or species in question can be used 
in developing disease-resistent varieties or population, once resistance-conferring genes 
have been found (Allard 1964). However, it must be kept in mind that disease 

jl/ Based on a paper presented at the Training Course in Forest Tree Breeding held in 
~ Canberra, Australia, 1977. 

- 179 - 

resistance involves two genetic systems, that of the host and that of the pathogen, each 
a result of the two basic variables of genotype and environment. The phenotypic ex* 
pression of a disease will thus depend on the interactions between the host, the patho- 
gen and the environment (Mclntosh 1971). As the pathogen populations constantly react 
to changes introduced into the host populations (Dinoor 1975), resistance breeding is a 
dynamic rather than a static process. A full understanding of the pest complex, including 
the variability and genetic potential of the host, the biology of the pathogen and the 
interactions between host, pathogen and environment, is a necessary basis for a program 
on breeding for resistance. 


Diseases in forest trees can be grouped into two categories, those caused by 
obligate and those caused by facultative parasites. Obligate parasites are restricted 
to living tissue, whereas facultative parasites are able to colonise both living and 
dead organic matter (Tarr 1972). The rust fungi are examples of obligate parasites in 
forest trees (e.g. blister rust of pines and leaf rust of poplars), while facultative 
parasites, which often are limited to certain provenances, include blight diseases (e.g. 
needle cast disease of pines and stem canker of poplars) (Bjorkman 1966). The suscepti- 
bility of trees to facultative parasites is generally greatly affected by their 
physiological condition and growth vigour, and thus indirectly by the environment 
(Schreiner 1966). 

Pathogenic organisms have enormous potential for developing new, virulent forms 
or races. Plant breeders must therefore be prepared to face new races of pathogen to 
which their formerly resistant varieties or populations are susceptible (Allard 1964). 

Once a fungal race becomes established, its prevalence is determined by the 
varieties of the host plants being grown. An extreme case is the growing of a crop con- 
sisting of inbred lines or one single clone in which the individuals are genetically 
identical, possessing identical genes for resistance against prevailing races of the 
pathogen. A host population like this will exert a strong selection pressure, and is 
likely to give rise to highly specialised physiologic races of the pathogen which are 
able to overcome the resistance of the host (Borlaug 1966). 

The key to understanding variability in fungi, including variability in pathogenic 
capabilities, lies in their reproductive systems. In addition to meiotic recombination 
of the genetic material, variation may be added to the population by non-meiotic recom- 
bination, by heterocayosis, in which two or more nuclei occupy the same cell, by cyto- 
plasmic inheritance, and by mutation which, itself, provides the initial genetic 
differences that are brought together in the various recombination processes (Buxton 1961). 
The majority of the species are haploid organisms in which nuclear fusions occur to give 
rise to a relatively short-lived diploid stage, but life-cycles ranging from completely 
haploid to completely diploid are also encountered (Allard 1964). 


Variability of the host leading to resistance can be either active or passive. 
Passive resistance is accomplished by structural barriers of a morphological or anatomical 
nature. Active resistance is due to vital processes initiated in the host as a reaction 
against a pathogenic agent, and is directed against the agent itself (G&uman 1950). 

Schrefner (1966) lists the following host characteristics as the most common ones 
contributing to disease resistance in forest trees: 

(1) Variation of the host in the uptake of minerals under similar field conditions; 

(2) Differences in water content of the host tissue; 

(3) The presence of metabolic products in the host tissue other than phytotoxins; 

(4) The presence of fungistatic or phytotoxic substances in the host; 

- 180 - 

(5) The production in the host tissue of phytotoxins; 

(6) Pre-existence of formation of barriers that restrict the movement of the 
pathogen into and within the host tissue; 

(7) Effect of osroosed substances on germ tube growth and penetration of stomata; 

(8) Escape mechanisms such as drooping needles or densely pubescent leaf surfaces. 

The genetic nature of host resistance can either be vertical (specific resistance) 
or horizontal (non-specific resistance, also called field resistance or tolerance) 
(van der Plank 1963). Vertical resistance acts against specific races or varieties of the 
pathogen; it is generally characterised by a close correspondence between the genes of the 
pathogen and those of the host, and is often controlled by single genes, although many genes 
may be involved. The genetic components that contribute to vertical resistance can 
generally be isolated and studied in appropriate crosses (Schreiner 1966; Watson 1971). 

Horizontal resistance is generally polygenic and many factors, such as control of 
rate of growth of the hyphae through the host tissue and penetration through the tissue, 
may simultaneously control resistance. Horizontal resistance generally gives a certain 
amount of protection against a whole range of races of the pathogen (Watson 1971). No 
direct correspondence exists between host and fungal genes, and although generally more 
efficient against fungi and favoured both in agriculture and forestry, polygenic systems 
are more difficult to utilise in breeding than the major genes controlling vertical 
resistance (Heybroek 1969; Luig 1971; Watson 1971). 

Present evidence for forest trees indicates that resistance may be due to a small 
nunber of major genes (e.g. in blight resistance of poplar hybrids; see Heimburger 1966), 
monogenic resistance (e.g. in resistance of western red cedar to leaf blight; see Soegaard 
1966), polygenic resistance (e.g. in resistance of western white pine to blister rust; 
see Bingham t al^. 1960), and cytoplasmic inheritance (e.g. in resistance of larch to 
needle blight; see Lagner 1952). Both major genes and polygenes may simultaneously con- 
tribute to resistance (e.g. in resistance of eastern white pine to blister rust; see 
Heimburgher 1962). 

While it has been generally accepted that different genetic systems operate in the 
case of vertical and horizontal resistance to a single disease, the same genes may be 
involved in both types of resistance (Watson 1966). A diversity of resistance genes in 
the population will reduce the chances of the pathogen overcoming existing resistance by 
a change in virulence or pathogenicity (Borlaug 1966; Watson 1971; Dinoor 1975). 

In addition to being genetically controlled, horizontal resistance may arise from 
imperfect synchronisation of pathogen and host, as in diseases which appear late in the 
life of the plant or late in the growing season (e.g. poplar leaf rust). This verges on 
disease escape, in which vigorous individuals may be able to replace diseased leaves or 
other parts, and so produce a reasonable yield in spite of infection. Plants may also 
escape infection if their susceptible phase happens to come at a time when the pathogen 
is absent, present in insufficient amounts, or when environmental conditions are un- 
favourable to it (Tarr J972). 

Assessing the control alternatives 

Before embarking on a program to breed pest-resistant trees, alternative methods 
of disease control should be carefully assessed. These include: (1) exclusion, 
(2) avoidance, (3) eradication, and (4) protection through chemical or biological control 
(McNabb 1964; Gibson 1975). 

Exclusion through quarantine, i.e. restricting the introduction of plants, parts 
of plants or soil into the country (as e.g. in Australia) may be effective against 
diseases not yet established, provided there are adequate means of enforcing the legis- 
lation. Exclusion through inspection and treatment of planting material and possibly 
tools and machinery before transferring them from one region to another pre-suppotes 
good information on the spread and life-cycle of the pathogen (Gibson 1975). 

- 181 - 

Disease avoidance through a change of species is usually effective if an accept- 
able alternative species is available (e.g. the change from one species of cypress to 
another in East Africa discussed later in this paper; the change from spruce to pine or 
larch on some sites infested with root rot in Northern Europe; see Palmberg 1969). 
Adjustment of e.g. sowing and planting times to reduce the risk of infection can some- 
times be successful, as can a change in management practices (e.g. a change in pruning 
routines of cypress in East Africa; see Rudd-Jones 1954). 

Eradication of other than nursery diseases must be achieved at a very early stage 
of the disease outbreak, and calls for an exceptionally efficient survey organization to 
be effective, e.g. eradication of diseased individuals of elm to stop the spread of dutch 
elm disease have largely failed (Neely 1975). Eradication of the secondary host in rust 
diseases has, however, proved fairly effective (e.g. removing aspen in Northern Europe and 
Ribes spp. in USA from areas close to pine plantations to avoid damage by pine twist and 
fusiform rusts; Day 1972). 

Protection by the application of fungicides is generally regarded as too costly 
outside the nursery. However, in diseases which attack the trees during a limited period 
of their life this approach can sometimes be economically acceptable (e.g. spraying of 
radiata pine against needle blight in New Zealand; see Gilmour and Vanner 1971). 
Protection by the application of insecticides is usually considered feasible during out- 
breaks of insect epidemics (Benedict 1964). Biological control has been very success- 
fully employed against insects (Franz 1964) (e.g. control of the Si rex wasp in Australia; 
see Anon. 1974) . 

The above control methods are in many cases practiced parallel with programs on 
breeding for disease resistance, and some of them provide only short-term solutions to 
a problem. Generally the most economical and, in many respects, ideal method of 
achieving disease control is the development of resistant biotypes of the host (Gibson 

Developing methods for evaluation of resistance 

Once a decision to breed for resistance has been reached, the development of rapid 
and accurate methods for evaluating resistance in the host plant is the most important 
immediate problem. Resistance is not an intrinsic character, but is subject to environ- 
mental influences on the host, the pathogen and the host/pathogen relationship (Schreiner 
1966). In addition to basic research on the establishment and progress of parasitism, 
the development of such evaluation methods will therefore require knowledge of the effects 
of environmental factors on host resistance and of the pathogenicity and virulence of 
the parasite (Callaham t ad. 1966; Schreiner 1966). 

As expression of resistance is phenotypic, the same expression may be the outcome 
of interaction between many different host and pathogen genotypes (Dinoor 1975). The 
identification of genes for resistance therefore involves crosses and genetic analyses, 
which in turn require the development of efficient techniques for selecting, propagating, 
breeding and progeny testing the host (Schreiner 1966). 

Identifiable resistance genes depend on the pathogenic strains present; in the 
absence of the pathogen, resistant genotypes are indistinguishable from non-resistant 
(Allard 1964). Programs for identifying resistance should therefore always be based on 
artificial inoculation (Allard 1964; Dyson 1974). As immunity to a disease is usually 
an unrealistic objective, it is necessary to set up minimum practical standards of 
resistance for each host/pathogen relationship. For both economic and biological reasons 
these standards must be kept flexible (Schreiner 1966). 

Selecting for disease resistance 

Aft^r developing methods for recognising and identifying disease resistance, the 
next step is to locate resistant material in existing populations of the host (Allard 

Resistance to disease in plants is more common than susceptibility, most wild 
plants being resistant to ttost diseases (Cowling 1969). Our most disastrous forest tree 
diseases have resulted from the introduction of foreign pathogens into populations which 
carry no effective genes for resistance to this pathogen (e.g. chestnut blight, dutch elm 
disease, white pine blister rust) (Borlaug 1966). Widespread planting of fast-growing exotics 

- 182 - 

sometimes under unfavourable ecological conditions and of populations possessing a very 
narrow genetic base, also easily leads to phytopathological problems (e.g. needle blight 
in radiata pine in East Africa, poplar leaf rust in Australia) (Bjdrkman 1966). 

Genes- do not occur at random in populations (Qualset J975). Long association 
between pathogen and host iff likely to lead to the development of mutual tolerance , with 
consequent elimination of highly susceptible host genotypes through natural selection 
(Tarr J972). Genes for resistance to pathogens and pests are therefore likely to be 
found in highest frequency in regions where the plant in question has been grown on a 
large scale in the presence of the pathogen, or by an extensive search of natural 

A large number of individuals should be screened at high levels of artificial 
inoculation to augment any search for resistant individuals in the forest. Resistance 
teats must be designed in consideration of the biology of the host and the biology of 
the pest; for example, the age of the host at which infection and diagnosis are to be 
made must be known. The trials should be replicated to sample a range of environments. 
The design must allow for statistical procedures to deal with variation in host, pest 
and environment, and their interactions. Tests on clone x environment interactions 
should not be considered conclusive until 1/3 to 1/2 of the estimated rotation elapses 
(Callaham et al. 1966).. Screening for vertical resistance should be undertaken in each 
planting region separately, whereas screening for horizontal resistance often can be 
centralised. However, factors like climate and the amount of virulent inoculum present 
in various planting regions may alter the level of resistance found at a central testing 
site (Dinoor 1975). 

If vegetative propagation is feasible for mass -product ion of planting material 
(as e.g. in poplars) and if the resistant trees have desirable phenotypes, they can be 
used immediately to develop pest resistant test populations. However, in the case of 
species reproduced by seed, it should be remembered that trees not only vary in resist- 
ance to a pathogen, but also in their ability to transmit resistance to their progeny 
(Wood 1966). It is therefore essential that phenotypically selected trees of these 
species are progeny tested. If resistant progenies emerge, their parents can tentatively 
be designated 'resistance transmitters* (Schreiner 1966), and after further tests pro- 
pagated into orchards to produce resistant Fj seed (Callaham e aU J966). Trees ex- 
hibiting high general combining ability for resistance will be of greatest use. Specific 
combining ability can be utilised only through vegetative propagation or through controlled 

Breeding for disease resistance 

Variation in the degree of resistance has been attributed to the kind and number 
of resistance genes present in the parent (Wood 1966). Breeding can be used to ensure 
that cultivars possess maximum genetic diversity for resistance genes (Watson 1966). 

In polygenically controlled systems a selection plateau may be reached beyond 
which further advances are insignificant. This plateau may or may not confer a degree 
of resistance adequate for practical purposes. Introducing new variation through hybridi- 
zation of the most resistant individuals and additional populations of the host which 
show some degree of resistance may overcome this plateau (Heimburger 1962). 

If sufficient disease resistance cannot be found in existing populations of the 
host species, genes for resistance can sometimes be introduced through inter-specific 
hybridization. If the resistant material thus developed is not commercially acceptable, 
two procedures are available: the Fj hybrids can be crossed to produce an F generation 
where segregation and recombination may result in acceptable resistant phenotypes, or 
the Fj hybrid can be backcrossed to the desired parent species (Allard 1964). Renewed 
screening for disease resistance must be undertaken in these new populations. 

Mass-producing inter-specific Fj seed possessing both desirable phenotypic charac- 
teristics and disease resistance may be done either through controlled pollinations or, 
where the two species involved flower simultaneously, they may be interplanted in 
orchards to produce hybrid seed spontaneously. In some cases single clones of one species 
can be interplanted within stands of the other species; this procedure facilitates roguing 
of non-hybrid seedlings in the nursery (Callaham e al. 1966). 

- 183 - 

Both intra- and inter-specific hybridization nay be used to transfer genes for 
resistance to desirable phenotypes. 

Maintaining resistance in improved populations 

There is evidence of sufficient variation in resistance between species, races 
and individuals to warrant the use of resistance breeding to combat practically all 
important plant diseases (Schreiner 1966). Whenever satisfactory disease resistant 
varieties are available they have been preferred over other means of control, because 
once developed they add little or nothing to the cost of production (Allard 1964). In 
crop plants which to a great extent consist of genetically homogeneous planting material 
and inbred lines, resistance breeding demands continuous study and mobilisation of new 
and different sources of resistance (Dinoor 1975). For example, the average maximum 
duration of effective protection by any given type of stem rust resistance in wheat in 
USA and Canada has been about 15 years; the situation has been even worse in respect of 
areas of the sub-tropics and the tropics, where the disease escape mechanism does not 
function effectively and where the gene pools for pathogenicity are broad and persist 
from year to year (Borlaug 1966). 

As long rotations are the rule in forestry and effects of any long-term climatic 
changes on the host, the pathogen and their inter-relationships may be pronounced, the 
question of maintaining resistance is of vital importance (Schreiner 1966; Painter 1966). 

The probability of attaining and maintaining resistance by breeding for extended 
periods of time is generally directly proportional to the diversity of germplasm avail- 
able to the breeder (Painter 1966). Whereas the basic unit in crop plant breeding is an 
inbred line or a strain of the plant, the basic unit in forestry is generally the indi- 
vidual, and an infinite number of genotypes is involved in each population (Painter 1966). 
As long as the genetic base in forest plantations is not allowed to become too narrow 
through vigorous selection in limited populations and through inbreeding, problems are 
unlikely to develop on the same scale as in agriculture. Broadly based genetic reserves 
may be an additional safeguard for the future; by this means the capacity of the host 
populations to react to the changes in pathogen populations may be preserved (Dinoor 


Although breeding for disease resistance in forest trees is a long-term project, 
success stories are numerous. 

Diseases attacking the leaves, limbs and stems of trees have been more thoroughly 
investigated than e.g. root rots. Although root rots are economically important, evalua- 
tion of resistance, screening of selections and progenies, and the development of effec- 
tive inoculation methods cause difficulties when dealing with internal damage. In addi- 
tion, root rot fungi usually invade the dead heartwood of the tree, in which the possi- 
bilities of introducing changes leading to resistance seem smaller than in living 
tissue. There have, however, been some positive results from altering by selection the 
toxic substances present in the heartwood (Bjorkman 1966; Cech ejt l. , 1966). 

Two examples, one entailing a change of species and subsequent selection, the 
other selection and further breeding, may serve to illustrate how the techniques on 
breeding for disease resistance outlined above can be successfully applied in practice. 

Cupressus macrocarpa Hart, and . Lusitanica Mill were introduced into East Africa 
in the early J900s. C. macrocarpa wag favoured in plantations because of its somewhat 
faster growth. However, problems were soon encountered with cypress canker, caused by 
the fungus Monochaetia unicornis (Cook & Ellis) Sacc. The pathogen was studied inten- 
sively in the early 1950s. It was found that it occurred on the indigenous species 
Juniperus procera Hook, in which it caused little damage. Three strains of the fungus 
were isolated, two unimportant and one virulent one; these were brought into culture to 
facilitate their study. Then an effective inoculation technique was developed, involving 
wounding the stem before the application of virulent inoculum. It was found that disease 
resistance of individual trees could be evaluated by measuring the rate of diameter 
increase of bark lesions on the trees over a period of three months after inoculation 
Using this method it was demonstrated that C. lusitanica was only half as susceptible 

- 184 - 

* B C inacrocarpa, and a change of species was therefore recommended and undertaken. The 
sane inoculation technique was subsequently used to test open pollinated C. lusitanica 
progenies of select plus trees (Dyson 1974). Some 80Z of the total reafforestation 
area in Kenya is presently planted with C. lusitanica. All seed used originates in seed 
stands selected for good form, high yield and a high degree of disease resistance. 

Western white pine (pinus nonticola Dougl.) is among the most valuable timber trees 
in USA. Blister rust, Cronartium ribicola Fisch. y was first accidentally introduced into 
USA in the 1920s, and by 1941 the disease had reached epidemic proportions. Observations 
in natural populations of the pine showed that variation existed in disease resistance. 
Apparently resistant or immune p he no types were selected, and resistance was tested by the 
response of grafts and progeny of selected trees over several years and at several sites. 
As the alternate host for the disease is Ribes spp. this plant was interplanted in the 
experimental plots to augment inoculum potential. Results from grafts and progeny trials 
indicated that both additive and non-additive gene effects were involved in resistance. 
The original selections were screened for high general combining ability by means of these 

The next step in the breeding program involved controlled crosses between promising 
phenotypes; these were tested for two years under intense artificial inoculation, using 
an "inoculation tent" which maintained humidity at levels favourable to the pathogen. 
It was found that no selection yielded progeny that were completely immune to the 
disease, and that a wide range of variation was apparent in the level of resistance 
transmitted by the selected trees (Bingham et l. I960; Hoff J966; Bingham J969). 

Practical breeding programs now utilise natural intra-specific variation in western 
white pine. The Forest Service of the Northern Region has embarked on a project to mass- 
produce ?2 seed from screened intra-specific Fj hybrids of pheno typically resistant 
parents. Heritability studies have indicated a gain of some 20Z of F] stock in disease 
resistance under intense artificial inoculation conditions. Practical problems encoun- 
tered involve seed orchard technology, size of gene pool needed for each planting area 
to assure adequate genetic variation for further improvement of desirable traits, and 
procedures to speed up selection and testing of new individuals (Hoff 1966). 

During the last ten years there has been a growing interest of inter-specific 
hybridization in the white pine group. There are more than 20 species of 5-needled white 
pines exhibiting varying degrees of resistance to blister rust. The breeding programs 
have concentrated on those 14 species which possess the widest adaption to varying site 
and climatic conditions and the strongest species-wide resistance. Comparative world- 
wide observations have been made on the relative resistance of the various species, and 
an international list of 'tentative rankings 1 has been drawn up. Resistant hybrids 
between a number of species have been produced and tested on an experimental scale; 
especially some of the Asiatic species (P. griff ithii, ]?. Armandii, . Koraiensis)and 
their hybrids have proved promising, exhibiting a much higher degree of resistance than 
e.g. P. monticola (Bingham J972). 

Steps in developing resistance through inter-specific crosses parallel those pro- 
ducing resistant strains within a species (Callaham e_t aJU 4966). 

The role of forest genetics will become progressively greater as forestry becomes 
more intensive. The growing of big, uniform plantations with stock often bred for non- 
fitness traits like growth and stem straightness in the absence of potential pathogens 
will increase the disease hazard. Efficient plant breeding programs should include plans 
to deal with the threat of undetected pathogens and new virulent races of existing ones 
before they can cause excessive damage. As stressed earlier in the paper, this can best 
be done by maintaining heterozygosity and a broad genetic base in the plantations as 
well as in the genetic reserves established or maintained in connection with the breeding 
programs . 

International co-operation is essential for rapid progress in disease resistance 
breeding. The co-operation should include genetic conservation of natural stands to serve 
as gene pools for disease resistance, international gene banks of pest resistant geno- 
types for the use in breeding, and an increasing exchange of information. 

- 185 - 

Allard, R.W., 1964. 'Principles of Plant Breeding 1 . Wiley & Sons Inc., New York. 

Anon., 1974. Commonwealth Scientific and Industrial Research Organization (CSIRO). 
Annual Report 1973-74. 

Benedict, W.V., 1964. Principles, procedures and problems in controlling forest pests. 
FAO/IUFRO Symposium on Internationally Dangerous Forest Diseases and Insects, Oxford 
FAO/FORPEST - 64, IX. FAO, Rome. 

Bingham, R.T., 1969. Rust resistance in conifers - present status, future needs. Second 
World Consultation on Forest Tree Breeding, Washington. FO-FTB-69, 5/2, FAO, Rome. 

Bingham, R.T., Squillace, A.E. and Wright, J.W., 1960. Breeding blister rust resistant 
western white pine. Silvae Genetica 9: 33-41. 

Bingham, R.T., 1972. Taxonomy, crossability, and relative blister rust resistance of 
5-needled white pines. 'Biology of Rust Resistance in Forest Trees 1 . U.S. Department 
of Agriculture, Forest Service. Miscellaneous Publication 1221. pp. 271-280. 

Bjttrkman, E. (1966). Status and trends in research related to the resistance of forest 
trees to disease in northern Europe, Gerhold, H.D. e al . (Eds). 'Breeding Pest 
Resistant Trees'. Pergamon Press, Oxford, pp. 3-JO. 

Borlaug, N.E. (1966). Basic concepts which influence the choice of methods for use in 
breeding for disease resistance in cross-pollinated and self-pollinated crop plants. 
Gerhold, H.D. t l. (EdsX 'Breeding Pest Resistant Trees 1 . Pergamon Press, Oxford, 
pp. 327-344. 

Buxton, E.W, (1961). Mechanisms of variation in the pathogenicity of Fusarium oxysporum. 
Recent Advances in Botany 1; 502-507. 

Callham, R.Z. et_ al . (1966). General guidelines for practical programs toward pest- 
resistant trees. Gerhold, H.D. et al. (Eds). 'Breeding Pest Resistant Trees'. Pergamon 
Press, Oxford, pp. 489-493. 

Cech, F.C. e al^. (1966). Breeding conifers for resistance to Fomes annosus. Gerhold, H.D. 
e l . (Eds). 'Breeding Pest Resistant Trees'. Pergamon Press, Oxford, pp. 483. 

Cowling, E.G. (1969). Principles of genetic improvement in disease resistance of forest 
trees. FAO/N.C. State Univ. Forest Tree Improvement Training Center. School of Forest 
Resources, N.C.S.U., Raleigh. Lecture Notes, pp. 196-200. 

Day, P.R. (1972). The genetics of rust fungi. 'Biology of Rust Resistance in Forest Trees'. 
U.S. Department of Agriculture, Forest Service. Miscellaneous Publication 1221. pp. 3-17 

Dinoor, A. (1975). Evaluation of sources of disease resistance. Frankel, O.H., and 

Hawkes, J.G. (Eds). 'Crop Genetic Resources for Today and Tomorrow'. IBP 2. Cambridge 
University Press, pp. 201-210. 

Dyson, W.G. (1974). Breeding for disease resistance. Report on the FAO/DANIDA Training 
Course on Forest Tree Improvement, Kenya. FAO/DEN/TF 112. FAO, Rome. pp. 292-298. 

Franz, J.M. (1964). Forest insect control by biological measures. FAO/IUFRO Symposium on 
Internationally Dangerous Forest Diseases and Insects, Oxford. FAO/FORPEST-64, IX,, FAO 

Gibson, I.A.S. (1975). 'Diseases of Forest Trees Widely planted as Exotics in the Tropics 
and Southern Hemisphere 1 . Commonwealth Mycology Institute /Commonwealth Forestry 
Institute, Oxford. 

- 186 - 

Gilaour, J.W., and Vanner, A.L. (1971). Radiata pine needle blight (Dothistroma pini) . 
Fungicide and Nematocide Tests 27. Am. Phytopath. Soc. 

GSumann, E. (1950) . 'Principles of Plant Infection*. Crosby Lockwood & Son, London. 

Heimburger, C. (1962). Breeding for disease resistance in forest trees. For. Chron. 

Heimburger, C. (1966). Susceptibility to a serious fungus attack as a genetic barrier 
between aspen species. Gerhold, H.D. e jalU (Eds.) 'Breeding Pest Resistant Trees' 
Pergamon Press, Oxford, pp. 391-394. 

Heybroek, H.M. (1969). Three aspects of breeding trees for disease resistance. Second 
World Consultation on Forest Tree Breeding, Washington. FO-FTB-69, 5/4. FAO, Rome. 

Hoff, R.J. (J969). Blister rust resistance in western white pine. Gerhold, H.D. et al. 
(Eds). 'Breeding Pest Resistant Trees'. Pergamon Press, Oxford, pp. 119-124. 

Lagner, W. (1952). Reziprok unterschiedliches Verhalten von Larchenbastarden gegen eine 
Nadelerkrankung. *L. Forst genet ik ^: 78-81 

Luig, N.H. (1971). 'Strong' and 'weak' genes for stem rust resistance in wheat breeding. 
Aust. PI. Breeding Conf . Perth, W.A. 8-3. 

Mclntosh, R.A. (1971). Wheat genes for rust resistance in space and time. Aust. PI. 
Breeding Conf. Perth, W.A. 8-5. 

McNabb, H.S. Jr (1964). A 'new* concept of forest tree disease control: physiological 
suppression. FAO/IUFRO Symposium on Internationally Dangerous Forest Diseases and 
Insects, Oxford. FAO/FORPEST-64, IX. FAO, Rome. 

Neely, D. (1975). Sanitation and dutch elm disease. Burdekin, D.A., and Heybroek, H.M. 
(Eds). 'Dutch Elm Disease'. Proceedings of IUFRO Conference, Minneapolis 1973. USDA 
FS Northeastern Forest Experiment Station, Upper Darby, PA. pp. 76-87. 

Painter, R.H. (1966). Lessons to be learned from past experience in breeding plants for 
insect resistance. Gerhold, H.D. t al (Eds). 'Breeding Pest Resistant Trees'. 
Pergamon Press, Oxford, pp. 349-355. 

Palmberg, C. (1969). Femes anno s us Fr. (Cke.) - a universal problem. Silva Fennica 

Qualset, C.O. (1975). Sampling germplasm in a center of diversity: an example of disease 
resistance in Ethiopian barley. Frankel, O.H., and Hawkes, J.G. (Eds). 'Crop Genetic 
Resources for Today and Tomorrow'. IBP 2. Cambridge University Press, pp. 81-98. 

Rudd-Jones, D. (1954). Studies on a canker disease of cypress in East Africa. Ann. App. 
Biol. 4J: 325-335. 

Schreiner, E.J. (J&66). Future needs for maximum progress in genetic improvement of 

disease resistance in forest trees. Gerhold, H.D. e jal_. (Eds). 'Breeding Pest Resistant 
Trees'. Pergamon Press, Oxford, pp. 455-466. 

Scegaard, B. (1966). Variation and inheritance of resistance to attack by Didymascella 
thu j ina in western red cedar and related species. Gerhold, H.D. et al. (Eds). 
' Breeding Pest Resistant Trees'. Pergamon Press, Oxford, pp. 33-38. 

Tarr, S.A.J. (1972). 'The Principles of Plant Pathology' Macmillan Press, London. 

Van der Plank (J963). 'Plant Diseases: Epidemics and Control'. Academic Press, New York, 
London . 

Watson, I.A. (1971). Breeding for disease resistance. Aust. PI. Breeding Conf. Perth, 
W.A. 8-37. 

Wood, F.A. (1966). The current status of basic knowledge of forest tree disease resistance 
research. Gerhold, H.D. l (Eds). 'Breeding Pest Resistant Trees'. Pergamon Press, 
Oxford, pp. 293-300. 

- 187 - 


B. Ditlevsen 
National Forestry Service, Denmark 



Introduction 187 

Market factors 187 

Choice of procedure 188 

Evaluation criteria for a genetic improvement plan 189 

Costs and benefits 189 

Costs 190 

Fixed costs 190 

Variable costs 191 

Benefits 191 

Dubious points in evaluation 192 

Optimization of improvement programmes 193 

Final comments 194 

Bibliography 195 


Tree breeding through genetic manipulation is one of several methods available for 
increasing forest production and the efficiency of processing procedures. The potential 
benefits and costs of tree breeding must therefore be considered within the overall frame- 
work of the task of reforestation and utilization as a whole. 

Restocking programmes may be promoted by individuals, private companies, ministries 
or other public bodies. If the capital comes from private sources, the usual objective 
is to maximize annual benefits. Projects promoted by public bodies may be established for 
social purposes, but also to obtain direct economic benefits. In the latter case, the 
objective will be to achieve maximum production rather than to maximize the economic 
benefits, in order to ensure that the country has an adequate supply of wood, or to deve- 
lop (or maintain) a strong forest products industry. 

In this paper we shall deal only with considerations relating to financial cost- 
benefit models, but the breeder must bear in mind that an analysis of social costs and 
benefits may be necessary in certain cases, and may lead to entirely different conclusions 
(Reilly, 1977). 


The market for genetically improved seed is like that of other commodities. Growers 
supply seed in accordance with existing demand. 

The demand for genetically improved seed depends on the nature and level of the 
demand for the commodities grown from the seed, and this, in turn, reflects the demand for 

- 188 - 

products made from these. This is of great importance, since the price that the buyers 
are willing to pay for the improved seed will be equal to the price of the products grown 
and manufactured from this seed, less the costs of cultivation and processing. 

Since the benefits of a tree breeding programme will extend beyond the first few 
years, it is advisable to have some idea of future trends in the market for the processed 
products. The level of demand forecast will influence the amount of seed to be produced, 
while the nature of the demand will influence the characters to be selected and th intensity 
of selection. 

In a stable market situation in which supply and demand, and hence prices, do not 
vary, the point at which the supply of genetically improved seed is in balance with demand 
will be mirrored in the market price of the seed. If the amount available on the market 
is excessive and demand does not change, the price will fall, but will eventually return 
to its original level as supplies are reduced. An increase in the demand for improved seed 
owing to an increase in the annual planting rate will result in a rise in the price. 

This very simplified description of the market relates to the prices of a surplus 
offered to other seed growers. If the surplus seed production of the organization's 
programme represents a considerable proportion of the total market, the price will fall. 
This must be borne in mind when evaluating the seed. Likewise, if the organization decides 
to buy genetically improved seed from other growers instead of starting its own programme, 
the increased demand may be sufficient to provoke a considerable rise in prices. However, 
if the effect on supply and demand is not very great, no great changes in price will occur. 

Markets for genetically improved seed are sometimes highly specialized* Breeding 
programmes cover a wide range of species, selected characters and selection intensities. 
The market is often dominated by one or a few producers, and the price will be based, not 
on market forces, but on the cost of production plus a margin of profit. The cost of 
producing genetically improved seed on the basis of one's own programme may be very different 
from the market price, or the price that an organization was willing to pay for the seed. 


There are two kinds of alternative procedures or strategies: those arising from the 
decision on whether or not to undertake the programme, and those arising from having 
chosen the improvement programme. The strategies in the first category are as follows: 

1st to continue using unimproved seed; 

2nd to buy genetically improved seed from other sources; 

3rd to participate in a cooperative tree breeding programme; 

4th to undertake one's own genetic improvement programme. 

Before undertaking a comparison of these strategies, it is necessary to ascertain 
whether it is technically possible for the organization to undertake a genetic improvement 
programme. This will depend essentially on the existing demand for genetically improved 
seed, the resources available to the organization, the size of the programme proposed, and 
the objectives of the directives. The resources available include not only highly skilled 
breeders and other labour, but also capital, land and, above all, the size and nature of 
the population available for selection purposes. Research may also prove necessary to 
establish flowering models, vegetative propagation techniques and other operational proce- 
dures for seed orchards. 

The kind of genetic improvement programme will depend mainly on the available data 
on the heritability of important improvement characters, the additional genetic gain at 
different levels of selection intensity, the genetic relationship between various characters, 
etc. The size of the programme will depend on the organization's seed requirements and 
the demand for seed surpluses* 

Initially, most organizations will not have sufficiently detailed information to 
undertake big programmes, and analysis will be limited to a comparison of these progr 

- 189 - 

with one or another of the above alternatives* As the programme proceeds and more is known 
about the genetic improvement characters of the species, more advanced strategies may be 


Economic experts have not yet completely solved the problem of choosing the right 
economic criterion for comparing the benefits and costs of alternative procedures over a 
period of years, or for accepting an individual project. But most now agree that it is 
necessary to have some method of discounting to reduce benefits and costs of different 
dates to a common denominator at any given moment. A dollar is worth more to an organi- 
zation today than at a future date, because it can be used to produce additional income. 
The following two evaluation criteria are the ones most frequently used: 

1 The current net value of a project is defined as the difference between the costs 
and benefits attributed to the project actualized at the appropriate rate. If the actual 
net value is positive, the project can be considered economically feasible and may be 
accepted. If there is more than one project, preference should be given to the one which 
offers the highest positive actual net value. 

2. The rate of internal yield is the type of actualization which gives an actual net 
value equal to zero, i.e. the actualized value of the benefits is equal to the actualized 
value of the costs. An individual project may be accepted if the rate of internal yield 
proves higher than the minimum acceptable rate adopted for investment purposes. The choice 
between projects wil 1 be determined in favour of the project with the highest rate of 
internal yield. 

In some cases, the project that offers the highest actual net value does, not 
necessarily offer the highest rate of internal yield. In such a case, the criterion of 
the actual net value is the one that should be followed. 

The actual net value of a programme (CNV) may be expressed in algebraic form through 
the following equation: 

CNV = f2 (B i- - O/C 1 + O't where: 

t=0 C L 

B denotes the benefits generated by the programme in the year t, 

C t denotes the coasts incurred by the programme in the year t, 

i denotes the rate of interest for actualizing benefits and costs, expressed as a decimal. 

In calculating the rate of internal yield it is not necessary to have any predeter- 
mined rate for actualization purposes. In order to decide whether the programme should be 
accepted or not, one needs an estimate of the minimum rate of yield acceptable for the 
organization contemplating the investment. This rate will be equal to that which the 
organization should have earned in alternative investments, i.e. the in timely placement 
cost of its capital. 

Most privately-owned organizations expect new investments to generate as a minimum 
a rate of profit equal to the weighted average profits from other sources (capital shares 
and loans, undistributed profits and external investments), according to the relative 
proportions of these resources used by the organization. 


The simplest method of analysing costs and benefits is to limit the analysis to 
actual cash transactions. The costs will therefore include wages and salaries, the price 
of purchasing land and other fixed aspets, such as buildings, plant and equipment, and 
the costs of the raw material. Values must be assigned to the fixed assets and land at 
the end of the time horizon and included as benefits. 

- 190 - 

Two aspects of the cost /benefit ratio must be examined. At what point in the pro- 
duction process should the costs and benefits connected with the genetic improvement pro- 
gramme be evaluated? And how should they be planned over the years? 

The costs must be charged to the account at the time they occur. The benefits are 
usually evaluated at the time the seed is collected. All the costs relating to the selec- 
tion of parent trees and the establishment and management of the seed orchard are naturally 
included in the analysis, while those relating to progeny and provenance trials, hybridiza- 
tion programmes, etc., are neither partially nor fully attributed to the main improvement 
programme, and it may therefore be necessary to earmark funds for these items in some way. 
It is preferable that this be from profits generated by the different programmes. Some- 
times, however, it is necessary to adopt an arbitrary procedure based on the breeder's 
own assessment. 

The value of genetically improved seed is based on its price on the free market 
(if there is one) or the value of profits generated by sale of the products obtained from 
the seed using a method called residual value price. This implies 'deduction of the costs 
of cultivation, harvesting, processing and distribution from the price of an end (or 
intermediate) product derived from the seed. The prices of sawnwood and pulp are appro- 
priate starting points for an evaluation of the benefits of a forest tree breeding programme, 

Fixed costs 

As in any other form of technological change, the costs of tree breeding may be 
fixed, or may vary in accordance with the size of the seed harvest produced. They are 
fixed if they are not affected by the size of the harvest, and variable if they vary in 
direct accordance with it. 

Selection and progeny trials of parent tree for the seed orchard constitute the 
main sources of fixed costs connected with the management of seed orchards. This is 
because the number of phenotypes selected to be propagated in the orchard is normally 
independent of the size of the orchard. 

1 Cost of selection per parent tree 

The cost of selection per parent tree for a seed orchard depends on the selection 
intensity and the criteria or characters on which selection is based. Van Buijtenen and 
Saitta (1972) showed that the initial selection cost per tree tends to increase at a 
relatively more rapid rate as selection intensity increases. Per example, the cost of 
initial selection per tree at high selection differentials of 3 or 4 standard deviations 
above the average was almost five times greater than with relatively little deviation 
from the standard. 

The cost of initial selection also increases greatly in proportion to the number 
of characters to be selected. For a selection intensity of one percent and assuming that 
the characters are not correlated and that the costs of the search amount to $1 per 100 
trees studied, the selection cost would be as follows: 

Number of characters Number of trees studied Cost per parent tree 
to be selected per parent tree selected ($) 

1 100 1 

2 10 000 100 

3 1 000 000 10 000 

The selection cost increases 100 times for each additional character selected. 

- 191 - 

2 Progeny trials 

A progeny trial may represent one of the most expensive aspects of the genetic im- 
provement programme, and its costs are very difficult to figure. If progeny trials are 
designed to provide information for the purpose of roguing a first-generation seed orchard, 
the total cost of the trials will be assigned to this phase of the improvement programme. 
If they are also designed to provide the basis for a second-generation selection, a method 
is required for assigning the costs between the first and second generation orchards. 
The ideal would be to base cost assignment on the profits generated by each generation, 
but it may prove impossible to quantify these. In this case, a general factor is needed 
to cover most circumstances, such as the relative costs of selection. 

3 Other fixed costs 

The annual costs of administration and research constitute the other most important 
categories of fixed costs* The costs of administration may be assigned pro rata among the 
costs of wages, salaries and materials directly generated in each phase of an improvement 
programme, but it is more convenient to treat them as a fixed annual cost. 

If the research programme is not designed to benefit the improvement programme 
carried out in combination with it, the costs must be changed to the future improvement 
programmes on behalf of which they are designed, when these programmes are produced. 

Variable costs 

The remaining costs of establishing and managing a seed orchard are closely connected 
with its area or size. For any management system, they also depend directly on the seed 
yield. They include the costs of preparing the site, planting, vegetative propagation, 
fertilization, pruning, roguing, overall maintenance and protection. 

Establishment costs will vary according to the number of clones, the method of 
propagation and the initial density of the forest. The per hectare seed production of a 
species depends on the extent of site preparation, particularly the amount of fertilizer 
used, and the density of the forest. Keiding (1975) has given a detailed summary of the 
costs of establishing seed orchards. 

The larger the area of the seed orchard, the lower the costs per kilogramme of seed 
produced. However, a moment may come when the orchard is too large and diseconomies of 
size or scale become apparent, owing mainly to a disproportionate increase in the costs 
of administration, or perhaps because the collection of graft material is very expensive 
or more parent trees are required. 

Thus the production cost of seed in an orchard will follow the well-known U-shaped 
curve of production theory, i.e. it will fall at the beginning as the area of the orchard 
increases, and rise at the end as the fixed costs start to rise. Similar relations may 
be observed in individual operations, particularly between the average cost of fertilizers 
per kilogramme of seed produced and seed production per hectare; and between the spacing 
of ortets and seed production per hectare. 


Tree breeding can benefit forest production in four ways: first, by increasing the 
yield of wood per hectare ("yield effect 11 ); second, by inducing a rise in the prices of 
the different products harvested ("price effect 11 ); thirdly, by reducing costs ("cost re- 
duction effect"); and fourthly, by decreasing the economic rotation for the forest 
("rotation effect"). 

1 Yield effect 

This effect can be obtained by cultivating genetically improved trees which give 
a higher yield, or by reducing losses in yield by cultivating trees better adapted to the 
environment, e.g. trees more resistant to wind in areas with cyclones or trees resistant 
to diseases, insects, damage produced by frost or snow or drought* One of the problems 
in trying to increase the inherent speed of growth, however, is the relatively low heri- 
tabllity of this character. 

- 192 - 

2 Price effect 

If we assume that the price of the improved seed will include any economic surplus 
from harvesting and processing of the trees grown from the seed, the following factors 
will influence the value or price of the seed: 

a) Higher yield of processed articles. For sawn logs, longer trees, with straighter 
trunks and improved form, and less compression wood and twisted grain, provide 

a higher yield of saw timber per unit of roundwood volume. For pulpwood, a 
higher yield is obtained from trees with greater basic density, little compression 
wood and better fibre properties. A reduction in the variability of these pro- 
perties of the wood also produces a higher yield. 

b) Improvement in the class or quality of the processed articles. Higher classes 
of saw timber, with consequently greater value, can be achieved by reducing the 
size of the branches and the amount of twisted grain and compression wood, 
improving branch angle, etc. Higher classification of the pulp can be obtained 
by cultivating trees with appropriate fibre properties. 

c) Reduction of felling costs. These costs can be reduced through higher yields 
per hectare, larger average trunk size at time of felling, wider spacing at 

time of planting, and better branch characteristics (reduction in snedding time). 

d) Reduction of processing costs. For saw logs, straighter, larger trees of im- 
proved form and uniform size reduce sawing time; improved wood properties can 
reduce drying time in the sawmills and cooking and grinding time in the pulp 

3 Cost reduction effects 

Faster-growing, better quality trees enable forestry directors to space tree plantings 
more widely, which means fewer seedlings per hectare and easier access. Many operations 
are thus shortened. 

* Rotation effect 

Higher prices and yields and lower cultivation costs mean economic rotation can be 
considerably reduced. This also offers the advantage of a reduction in the amount of land 

The benefit side of the budget is the most difficult aspect in any genetic improve- 
ment programme. How can the benefits deriving from the programme be evaluated? As we 
have already indicated, evaluation can be done in two ways: one based on the market price 
of the seed produced, and the other based on the residual value price method. 

If the market price method is adopted - which is the exception rather than the rule, 
since the improved seed, even of the same species, is rarely of the same genetic quality - 
the problem is to identify any change in price deriving from the increase in the supply 
of seed. 

If the residual value method is adopted, the procedure used will depend on the 
products cultivated or processed from the seed of the seed orchard and the characters 

The benefits to a plantation from genetic improvement can be evaluated in terms of 
the seed produced, actualized to the time the seed was collected and from the time profits 
begin to accumulate, using data relating to the initial planting density, the gerrai native 
capacity of the seed and the rate of roguing in the nursery to express the benefits in 
terms of dollars per kilogramme of seed. 


Most studies imply that physical production inputs and outputs and their prices can 
be estimated without error. But this is not so. 

- 193 - 

An analysis which takes into account the doubts inherent in probability distribu- 
tions for basic inputs and outputs and their values is much more exact with regard to the 
collection and analysis of data than one based on average or probable values. Also, it 
is very often impossible to make reasonable subjective estimates of the probability distri- 
bution of many variables, particularly in the analysis of investments that imply techno- 
logical changes, such as the genetic improvement of trees, in which even the average increase 
in benefits is very often unknown. 

One possible way of resolving this difficulty is to assume that each option evaluated 
will have the worst possible result, and to select the one that gives the highest value 
for the economic criteria selected. This is a highly conservative method and may result 
in projects with high potential benefits not being considered. 

In projects with investments of a biological type the costs can usually be evaluated 
fairly accurately, but the benefits raise difficulties. Tree breeding is a project of 
this kind. Lundgren and* King (1965) and Davis (1967) resolved the problem by calculating 
the levels of genetic gain that should be obtained in a tree-breeding programme based on 
known costs and then comparing these gains with data available from progeny trials. Another 
method is to limit the studies to gains that can be evaluated easily, as, for example, in 
the analysis by Swoford and Smith (1971) of the national forests in southern U.S.A. 

Finally, it may prove more satisfactory to make a sensitivity analysis of the key 
parameters, such as prices, or of those which are the object of subjective evaluation or 
are subject to big fluctuations. This requires an evaluation of each project on the basis 
of the most probable floor and ceiling values of the key variables. The main objective 
of the analysis is to judge whether it is to be assumed that a project is unacceptable 
in terms of its assumed floor, or whether the order of the competing projects should be 


According to the traditional production theory, the optimum result of a project is 
defined by the point at which the marginal cost is equal to the price of the production, 
where the marginal cost is defined as the cost of producing an additional production unit. 
For a project with a life of one year or less, this is relatively easy; but for long 
projects like tree-breeding programmes, the situation is complicated by the additional 
variable represented by time. 

The simplest solution to this problem is to assume that there are no restrictions 
on resources and then submit all the feasible genetic improvement options to separate 
cost-benefit analysis. The one that generates the maximum economic benefits will be chosen. 
Options may vary according to the number and nature of the characters to be selected, 
selection intensity, the number of plus trees propagated in the seed orchard, its area and 
the management system adopted. For each set of characters to be selected, there will be 
a range of production possibilities based on the above factors. 

This method may prove extremely complicated and require much time as the number of 
feasible improvement programmes increases. Owing to inadequate data this very rarely 
happens; but if sufficient information is available, mathematical programming techniques 
may prove appropriate. 

Linear programming was used by van Buijtenen and Saitta (1972) to derive optimum 
solutions based on the size of the harvest, the area of the seed orchard and whether or 
not the orchard had to be rogued. The technique is very appropriate for the optimum ear- 
marking of scarce resources between competing requirements, in terms both of individual 
programmes and of time. The process can be performed rapidly and exactly by computers. 

A more recent development is Porterf ield's study (1974) on potential gains in genetic 
improvement programmes for Pinus taeda in the southern U.S.A. using goal programming. This 
is simply a modification of lineal programming, but here the restrictions are replaced by 
the goals. Many goals can be included in this technique, while there is only one in lineal 
programming, i.e. maximization of the net returns in a single objective function. Characters 
which it was considered would affect the volume and specific weight were incorporated in 
the model. Only characters affecting the volume were taken into consideration for sawlogs, 
but both characters were taken into account for pulpwood. 

- 194 - 

The main advantage of the model is that the goals can be modified, absolutely or 
relatively, according to changes in the market. It can provide for the genetic selection 
reaction of indigenous populations, roguing and progeny trials. By specifying the selection 
intensity for a given character and the percentage of improvement desired, a solution is 
obtained which gives the minimum deviation from the given genetic goal subject to restric- 
tion under the capital expense budget. 

9 The purposes of any optimization will depend on the organization contemplating the 
investment. In view of the rapid changes in the market for wood products, many organiza- 
tions prefer to adopt a conservative genetic improvement strategy which maintains maximum 
flexibility. Greater importance may be attributed to characters that may be important for 
various generations in a range of different environmental and economic conditions. As a 
general rule, an increase in the yield of wood per hectare and greater resistance to diseases 
are two characters of this kind, while it may be assumed that wood properties such as grain 
are not so important. 

Other organizations are more concerned with improving first-generation characters 
which are obviously inferior but which might provide large gains rapidly, and concentrating 
more on other characters in later generations. 

Great care must be taken in applying very advanced optimization procedures, since 
they depend entirely on the data on which they are based. The benefits or gains deriving 
from the tree breeding constitute a major problem. Very little is yet known of the gains 
obtained through genetic improvement of trees in a complete rotation of a plantation re- 
generated by genetically improved seed, particularly in cases in which other cultural 
techniques, such as pruning and thinning, are used. For plantations that produce only 
one product, e.g. pulpwood, have a short rotation and are not thinned, the gMns can be 
forecast with more accuracy than for those which are regularly thinned and which produce 
both logs and pulpwood in long rotations. 


Genetic improvement is an expensive programme that must be justified by its potential 

Nikles (1973) gives a series of examples of the economic considerations relating 
to the genetic improvement of forest trees. 

Breeders usually choose the improvement programme which provides the biggest and 
most rapid gains and has the shortest improvement cycle. 

Smith and Zobel (1974) have shown that there will be enormous losses in potential 
gains if the source of the seed is unknown, and it is important that this source be con- 
sidered and thoroughly investigated before a genetic improvement programme is formalized. 

They also considered that important gains in straight ness of trunks, specific 
weight, length of tracheid and resistance to rust can be confidently assumed, as they are 
highly inheritable. 

As regards costs, they proved that, when all the costs of a representative seed 
orchard were updated to a common point in time, the selection of parent trees with the 
subsequent progeny trials represented only 10 percent of the total costs. Hence an 
improvement in selection standards may produce only a relatively small increase in costs. 
This bears out the results of van Buijtenen and Saitta (1972), who showed that selection 
was the most effective phase in the tree-breeding programme they were studying, in terms 
of the profitability of investment in a first-generation seed orchard. Most (89 percent) 
of the total costs of the seed orchard programme concern preparation of the site, cost of 
the land, supervision, fertilizers, protection, harvesting and extraction of seed, and it 
is important to realize that these costs do not vary with the genetic quality of the tree 

Some characters may compete with each other. This seems to be characteristic of 
characters of quality, such as straightness of trunk and branches, when these are selected 
in conjunction with fast growth. 

- 195 - 

The profitability of tree breeding is considerably increased by progeny trials and 
roguing of the seed orchard (Smith and Zobel, 1974). Consequently, if it is desired to 
obtain maximum economic benefits, progeny trials and high selection standards must be used. 

The seed yield per hectare of the seed orchard must also be maximized through appro- 
priate use of fertilizers and proper spacing. 

Owing to the variety of species and the doubts regarding benefits, it is difficult 
to generalize about the most appropriate method of formulating a tree improvement programme. 
Perhaps all we can do is to repeat the conclusion of Smith and Zobel (1974) that the most 
profitable tree improvement programmes usually have the following characteristics: 

1 One species is grown over a wide area; 

2 The desired characters have moderate to high heritability; 

3 Crossing of the species can be easily controlled, and the species is a prolific 
seed producer; 

4 The values of the wood products are high and it is to be assumed will remain 


Davis, L.S., 1967. Investments in loblolly pine clonal seed orchards: production costs 
and economic potential. J^ For* 65. 

Keiding, H. , 1975. Economic Considerations in IMproved Seed Sources, FAO/DANIDA Training 
Course on Forest Seed Collection and Handling, Chiang Mai, Thailand, 1975. 

Lundgren A.L., and King, J.P., 1965. Estimating financial returns from forest tree 
improvement programs. Soc. Amer. For. Proc. 

Nikles, D.G. , 1973. Economic aspects of tree improvement. Cost benefit of tree-breeding 
programmes. FAO/DANIDA Training Course on Forest Tree Improvement, Kenya. 

Porterfield, R.L., 1974. Predicted and potential gains from tree improvement programs - 
a goal programming analysis of program efficiency. N.C. State Univ. Tech. Rep. 52, 
Raleigh, N.C. 

Reilly, J.J., 1977. Economic Aspects of a Tree Improvement Program, International Training 
Course in Forest Tree Breeding, Canberra, Australia, 1977. 

Smith H.D., & Zobel, B.J., 1974. Genetic gains and economic considerations. Tree Improve- 
ment Short Course, School of Forest Resources, N.C. State Univ., Raleigh, N.C. 

Swofford, T.F. & Smith, O.D., 1971. An economic evaluation of tree improvement in the 
southern national forests. U.S.D.A. For. Serv. South Reg. Publ. 10. 

Van Buijtenen, J.P. & Saitta, W.W. , 1972. Linear programming applied to the economic 
analysis of forest tree improvement, J. For. 70/3. 

- 196 - 

Participants visiting the CON ARE Pinus caribaea plantations in Chaguaramas. 

- 197 - 

Palmberg, Paul and Willan 



Introduction ................................................... 198 

Preliminary considerations ......................................... 198 

Administrative matters .......................................... 198 

Technical questions ............................................. 198 

Selection criteria and the basis for selection .......................... 199 

Selection criteria ............................................. 199 

Basis for selection ............................................ 199 

Strategy ...................................................... 200 

General principles ............................................. 200 

Populations .................................................. 200 

Considerations in choice of breeding methods ......................... 201 

Breeding methods and their application .................... ......... 202 

A dynamic breeding programme ....................................... 202 

Current seed requirements ....................................... 202 

Base for future selection ....................................... 203 

Progressive improvement of seed through successive orchards and clone banks ... 203 

Conservation of gene resources .................................... 203 

Demonstrating progress .......................................... 203 

Staff organization, facilities and administration ........................ 204 

Staff and facilities .......................................... 204 

Revisions of plans and publication of results ... . . ..... . ............. 204 

Summary ...................................................... 204 

Bibliography .................................................. 205 

Annexes 1-6 Strategy model for Improvement of specific species 

This lecture is a revision of the one originally presented by Or D.G. Nikles at the 
FAO/DANIDA Training Course in Kenya in 1973, and published in the Report of that 
Course (FAO/DEN/TF-112. FAO t Rome 1974). 

- 198 - 


Planning a tree improvement programme involves many considerations. A primary factor 
is the stage in development of the reforestation project. A common situation, and the one 
adopted for reference in this lecture, is where a species with considerable known or 
suspected geographic variation is already being successfully planted on a moderate scale. 
The aim of the reforestation project is to meet the needs for a variety of products for 
local use and/or export. The scale of operations is sufficient to justify a continuing, 
multi-generation improvement prog ran 

Many more and less advanced variations of this reference model can be readily 
envisaged, and details of appropriate associated tree-breeding programmes would vary 
accordingly. It is not possible to consider many such situations here, so only the 
principles of planning a tree improvement programme are emphasized in this lecture. 


The first step in planning a tree improvement programme is to ascertain the types 
of products likely to be required and the aims of forest management now and in the future. 

Such an investigation and consideration of the following points will be needed in 
order to develop a strategic plan for the tree improvement programme: at step two the 
following are examined. 

A. Administrative considerations 

(i) Formulation and statement of the aims of the tree improvement programme in 
order that they will contribute most effectively to the overall objectives. Aims must be 
expressed as simply and precisely as possible. 

(ii) Assurance of possibilities to provide funds, equipment, facilities, and 
qualified personnel. Provision for extra training may be required. 

(iii) Assurance of programme continuity. It is desirable that suitable personnel 
be encouraged to make a career of tree breeding. If there are doubts about continuity of 
competent staff, it is essential to choose simple, robust strategies for the programme. 

(iv) Possible reorganization of administrative resources to locate tree breeder 
and staff in suitable headquarters together with colleagues working on silviculture, soils 
and nutrition, wood quality and products, forest management, etc. 

(v) Consideration of cooperation in improvement work at local, regional, national 
and international levels. This can lead to cost sharing of technical assistance and 
research, exchange of ideas and generation of enthusiasm. 

B. Technical considerations 

(i) Determination of the factors limiting forest production in the region and 
ways and means of manipulating these to meet stated overall objectives. It is essential to 
build a tree improvement programme on a base of sound silviculture, management and utili- 
zation. The relatively long time scale of tree improvement work and the possibility of 
rapid technological change in other techniques of forest management and utilization should 
also be borne in mind. 

(ii) Choice of species and provenances to provide the desired type of products. 
Identification of the best species and provenances for each major site type in potential 
planting areas is essential. 

The availability of basic biological information on the species (i.e. their ecologi- 
cal and morphological variability, individual variation and ease of regeneration by seed 
and vegetative means), as well as of possible techniques for improvement based on 
experiences in other countries should be investigated. Special characteristics of indivi- 
dual species can often be used to advantage. 

- 199 - 

(ill) Determination of the characteristics which are best manipulated by genetic 
means* Basic studies to secure reliable estimates of genetic parameters will provide 
information of great value in improving efficiency of selection and of breeding strategy. 

The development of simple assessment methods, efficient record keeping and data 
handling procedures will often be required in connection with these studies. 

(iv) Specification of the number of generations and number of years needed to 
achieve a certain degree of improvement in important characteristics* This information should 
be critically examined in the light of urgency for the improved material. 

(v) The gathering, throughout the programme, of data for the evaluation of costs 
and benefits. 

Evaluation of costs and returns have been discussed in the literature i.e. by Van 
Buijtenen (1975); Porterf ield (1978) : Reilly and Nikles (1978); Reich and Carlisle (1978); 
and van der Meiden (1978). 

Selection Criteria 

In choosing characteristics to be included in the programme, one should concentrate 
on a few traits which have high potential for economic gain and which cannot be more 
cheaply improved by cultural or technological means. Consideration should be given to 
inclusion of a trait if one or more of the following factors are relevant. 

(i) economic value is high; 

(ii) the trait is likely to be of continuing value even if product demands change 
in the future. 

(iii) variability and heritability are high, giving potential for considerable 
genetic gain; 

(iv) the trait is positively correlated with, or is independent of, other 
desirable traits; 

(v) genetic and management and non-genetic improvement methods interact favourably, 
or combine positively to improve the trait. 

Some traits are desirable for virtually all end products, and they may be expected 
to retain their prime importance fora long time. These are good health, high growth rate, 
lack of malformation, straight stem, large branch angle. Several other traits are of great 
value for specific products, e.g. special wood properties, branch pattern, branch diameter, 
branch longevity, and stem defects such as cone holes. 

In planning a programme, attention must be given to ranking the possible selection 
criteria in order of likely economic importance many years hence, and adjusting this 
ranking for expected genetic improvement. Each tree selected as a candidate for the breeding 
programme can then be assessed for overall merit. 

The Basis for Selection 

The degree of genetic improvement in the desired traits through selection and breeding 
is dependent on three components. These are (i) variation; (ii) heritability and 
(iii) proportion of trees selected. The degree of improvement is expressed by the Genetic 
Gain, AG, which is computed as the heritability x selection differential. 

'(i) Variation in the traits to be improved is a prerequisite and a starting point 
for any programme. Initial selection should be done in large populations in which the genetic 
base has not been narrowed down by e.g. collecting the seed for the establishment of the 
stand from only a few trees. 

- 200 - 

(ii) Heritabtlity (h ) is a measure of the degree to which a character 
influenced by heredity as compared to environment, and is expressed as the proportion of 
genetic to genetic plus environmentally caused (phenotypic) variation (see paper on 
quantitative genetics for more detail)* If selection is carried out in even-aged, regularly- 
spaced stands growing on uniform sites, the masking effects of the environment will be 
minimised, and accuracy of determining the genetic proportion of the variation will increase. 
This accuracy can be further increased by stratifying the plantation into site class areas, 
by comparing candidate superior trees to neighbouring trees growing on similar sites, 
applying adjustments to total volume for crown size effects, etc. 

(iii) Proportion of Trees Selected. The gain component most amenable to influence 
by the breeder is the proportion of the population selected. The intensity of selection is 
measured by the selection differential (s), i.e. the difference between the mean of the 
selected trees for a certain characteristic and the mean of the original population for the 
same characteristic. In theory, the larger the selection differential the larger the genetic 
gain. In practice, however, to increase the selection differential becomes increasingly 
more costly as the area to be searched increases in dimension logarithmically for a given 
increase in selection differential (Shelbourne 1973). Thus, a costly search for the 'l-in~a- 
million tree 1 will not achieve a correspondingly large selection differential, and less 
rigorous selection of individuals for inclusion in the first-round breeding programme may 
be called for (say, 0,1-11). In practice, countries with rather small plantations will have 
to select even less intensively (probably about 1 percent of the trees). 


Determining the most efficient system of managing the various parts of a breeding 
programme within the constraints of time and other resources available can be complicated 
and compromises are necessary to resolve conflicts of interest. The scheme that is 
developed as a result of all the conditions necessary is termed the breeding strategy. The 
process by which the strategy is implemented is termed the breeding method, while the 
biological procedures such as grafting, controlled pollination, planting, etc., are the 
breeding techniques. Thus, a breeding strategy is an approach to securing the greatest 
improvement with the species, resources and conditions prevailing. 

General principles 

Tree improvement has three major objectives, namely, to provide: 
(i) genetically improved seed or other material for immediate use; 
(ii) suitable select breeding material for future use; 
(iii) adequate genetic information for immediate and future use. 

In addition, the need to conserve the gene resources of the original populations and 
land races of the species concerned must be recognized. 


The principal objective of the tree breeder is the generation of optimum breeding 
populations to create cumulatively better genetypes for advanced-generation seed orchards 
(Namkoong 1972). This involves the reduction of the genetic base of the population, thus 
creating a serious conflict between short-term and long-term gains through reducing 
effective populations site. The conflict can be mitigated by maintaining a hierarchy of 
separate populations side by side, representing a series of decreasing selection intensity 
but increasing effective population size (Burdon, Wilcox and Shelbourne 1978). These 

- 201 - 

populations include the following: 

(i) Gene pool - a population in which the full range of genetic variation is 
maintained; it may sometimes equal the original selection population. 

(it) Selection population (base population) - a large population within which 
superior trees are selected; about 1 million trees is suggested (500 - 
1 000 ha). 

(iii) Breeding population - a selected population, containing perhaps 200-300 trees 
chosen for their superiority within the selection population. It is used in 
whole or in part to generate the next selection population. 

(iv) Seed production population - a population of 30-100 trees established to 

produce seed for forest plantations (or, for the wood-producing population). 

There are many variations of this pattern. Mass vegetative propagation would elimi- 
nate the seed-producing population as a separate entity, while a seedling seed orchard 
combines (ii) and (iii). 

The classic seed-producing population is the clonal orchard, giving very high selec- 
tion intensity but still large enough, say 25 unrelated clones, to give genetic "insurance 1 *. 
Maximum outcrossing is desired in the seed-producing population although the parents could 
be inbred themselves. The breeding population, which would include the seed-production 
genotypes, represent a compromise between maximum selection differential and full population 
size, but would also allow recruitment of individuals which are superior for particular 
traits. A gene pool of standing trees, although existing for gene conservation, cannot be 
shielded from all selective processes. Mild and non-specialized silvicultural selection 
seems reasonable, while preservation in several contrasting environments should help 
maintain genetic diversity (Burdon, Shelbourne and Wilcox 1978). 

Considerations in choice of breeding methods 

Gains likely to be achieved through tree improvement can be substantial and will be 
of a lasting nature. If breeding populations are kept variable and reasonably large (never 
fewer than 50 unrelated individuals), steady progress can be made for many generations 
through selection. 

An understanding of the following biological principles and technical guidelines may 
facilitate decisions when carrying out the improvement programme: 

(i) Most individuals of most forest species being used in large reforestation 
projects suffer depression of growth upon inbreeding. It is therefore essential to maintain 
a large genetic base in the selection and breeding populations to allow for mating between 

(ii) With small selection populations, say, only 1 000 individuals, a sacrifice in 
gain will be necessary; in order to maintain a large enough population (i.e. selecting a 
large proportion of the trees), selection intensity will have to be relatively low. 

(iii) Some individuals may have a very high general combining ability (GCA), which 
can only be utilized through controlled crosses. Provision for controlled crosses at an 
experimental scale should therefore be made at a fairly early stage of the programme. 
Choice of appropriate mating designs is of fundamental importance for this step in the 

(iv) Establishment of early trials is necessary to enable estimation of the effects 
of family-by-environment interactions. 

(v) Family experiments should normally contain a large number of entries and of 
the order of 100 siblings per family. The number of trees per plot should be small and 
replications numerous. 

(vi) Exploitation of the biological characteristics of the species with ingenuity is 
part of the art of tree breeding, for example, by means of clonal propagation, mass 
poilination t hybridisation* etc. 

- 202 - 

(vii) Vhere breeding programmes lack a sufficient number of desirable clones the 
situation can be rectified by selecting locally, or importing, additional material for 
injection of "new blood" into the breeding population. 

Breeding methods and their application 

Essentially, there are only two main strategies: one using sexually reproduced 
progeny and the other, vegetatively reproduced progeny. 

(i) Selection with Regeneration through Seed 

(a) Simple or recurrent mass selection. Example: Selection of phenotypically 
superior trees, collection of open-pollinated seed and bulking of the seed. Seed from the 
original, select trees is harvested continually (simple mass selection), or selection and 
seed collection is eventually carried out within their progeny (recurrent mass selection). 
The method is cheap, and it can be quite effective where the selection differential is 
large and phenotypic selection is accurate (high heritability and environmental effects 
minimized within the selection population). It has been used successfully in many countries 
and for many species. 

(b) Selection and simple progeny testing, with or without further control of 
pollination. Example; Culled seedling seed orchards (pollination control achieved by 



isolating the progeny from major sources of contaminating pollen). 

(O Selection, full control of pollination, without progeny testing. Examples; 
Clonal and control-pollinated seedling seed orchards. 

(d) As above but with periodic progeny testing, recurrent roguing and subsequent 
selection. Examples; Clonal and control-pollinated seedling seed orchards (a) with 
roguing of clones or of families and individuals; or (b) using test results to select 
material for new orchards. Note that roguing clonal seed orchards may give limited gain 
unless they contain a relatively large number of clones (50+) and the orchards have 
originally been established using relatively close planting distances. In some cases, where 
specific combining ability has been found to be high, bi-clonal orchards may be established 
for seed production. 

(e) Hybridization of species or provenances. Examples; Larch species hybrids in 
Europe and Japan; P. elliottii and P. caribaea var. caribaea in Queensland. This method is 
attractive in some~special circumstances (Brown, 1972). 

(ii) Selection with Clonal Propagation; 

(a) Without testing. Example; Some poplar and willow culture; . radiata in 
New Zealand to a limited extent. ~ 

(b) With clonal testing. Example; Poplar and willow culture; Cryptomeria in 
Japan proposed for P. radiata for future in New Zealand and for Eucalyptus hybrids in Congo. 
This is a high gain but "dead end" procedure unless a crossing programme is included to 
produce new selection populations periodically. 


Current seed requirements. Possible sources of improved seed for immediate use 
include the following: 

(i) Phenotypic selection and seed collection in local stands of good provenance 
generally assures well-adapted populations and gives modest genetic improvement. Genetic 
gain is limited by the unselected nature of the pollen parents. 

(ii) Collection of seed from thinned seed production areas; ideally, these should 
be chosen in stands of superior provenance. 

- 203 - 

(ill) Imported seed. This may be from good phenotypes of suitable provenance in 
natural stands or excess seed from an external breeding programme using an appropriate 
provenance grown in similar conditions. Improved imported seed should only be used to 
enrich a locally selected gene pool, it can never sibstitute local selections. 

Base for future selection. At an early stage in the breeding programme, it is 
appropriate to begin the importation of select genes (as seeds, scions or pollen) to broaden 
and improve the genetic base available locally for second generation selection. There are 
several examples of success in this procedure: scions, open-pollinated seeds and pollen of 
some 85 select trees of Pinus pinaster imported into W. Australia in the mid-sixties; 
vigorous exchange of select genes of P. radiata among several countries; scions and seeds 
of many P. caribaea var. hondurensis select trees exported from Queensland, Australia, to 
enrich gene pools elsewhere; seeds from selected orchard clones and ortets of P. eliottii 
and P. taeda in the U.S.A. exported to several countries in the southern hemisphere. 

It is necessary to watch closely for diseases and insects in any exchanged material 
and especially in scions. Watch for potential undesirable hybridization when planting 
important breeding material; interspecific crossing has caused a lot of trouble with 
eucalypts in Brazil. 

Progressive improvement of seed through successive orchards and clone banks 

The above means for providing somewhat improved seed to meet immediate requirements 
should be implemented as soon as possible in the programme. However, such techniques give 
relatively small gains. Higher gains are achieved in classical seed orchards, although the 
time-lag is longer. 

(i) Production orchards. A compromise and a useful strategy to gain time, is the 
establishment of clonal (or seedling) seed orchards (annually using the best material 
available each year. 

(ii) Progeny populations. Controlled crossing is used for the production of 
families for second-generation selection. Information for roguing the orchards is obtained 
from progeny tests established soon after parental selection, as well as from the full-sib 
progeny studies in which selection is carried out. 

(iii) Breeding population in clone banks; As explained earlier the seed production 
population is continually reduced to a superior nucleus (never less than some 25 clones 
per production orchard); the roguing is based on progeny trials. Parallel with these 
orchards it is essential to create highly variable, high-quality new selection populations 
for successive generations of selection. This requires the mating or large numbers of 
unrelated trees. It can be accomplished by vegetatively propagating select trees and making 
the crosses in a clone bank established for this purpose. Thus, large diverse selection 
populations and relatively narrow seed production populations are continuosly generated and 
selected. Special measures may be required to maintain diversity in breeding and selection 
population such as importation of new material or production of "wide crosses' 1 (Zobel and 
Me El wee, 1964; Zobel et al^ 1972). If adequate precautions are taken to keep a "sufficient" 
number (>50) of unrelated individuals in the breeding population, one would expect to be 
able to make steady progress for many generations without needing to add comparatively 
unselected new stock simply to restore lost genetic variability. 

Conservation of gene resources. The tree breeder must be concerned with conservation 
of the diversity found in the natural populations, both for future research and as a source 
in the future of gene combinations not currently required (Kemp et al, 1972; Nikles 1973-b; 
Yeatman, 1973; Zobel, 1973). Strategies are discussed in the lecture on conservation, 

Demonstrating progress. Libby (1973) listed 'political demonstration' as useful forms 
of planting. These are conveniently located plantings of simple design, the purpose of 
which is to show the achievements of the programme in as unambiguous a manner as possible. 
It may be desirable to deliberately include poor material, e.g. unsatisfactory provinces 
as well as 'routine' controls and 'improved' material. 

- 204 - 


Staff and facilities; It is essential to appoint an interested, well-trained and 
capable professional tree breeder to plan and develop the breeding programme in consulta- 
tion with senior managerial and utilization personnel. For a large-scale programme, 
the tree breeder should be supported by well-trained technicians and labour of high 

At the headquarters, or within easy access, there should be biometrical services 
and library facilities. It may be necessary to take special steps to develop or obtain 
efficient techniques and computer programmes designed to facilitate the collection, 
processing and analysis of data. The tree breeder and the biometrics team should keep 
abreast of modern developments in mating and field designs and work on their application 
to local problems. Opportunity should be provided for the tree breeder to participate in 
regional, national and international scientific meetings, workshops and refresher courses 
from time to time. The local group or team of workers at headquarters should be encouraged 
to hold occasional seminars and discussions at which progress and problems being 
encountered in their programmes, each contributing to the overall aims of management, are 

Cooperation with other tree breeders working with the same species under similar 
conditions is most worthwhile, and periodic reciprocal visits by the breeders among 
programmes should be arranged. 

All these aspects help to ensure constant review of the direction and aims of the 
work and a vitality and sense of purpose in the personnel. 

Revisions of plans and publication of results; The tree breeder plan should be 
revised periodically when thorough re-examination of objectives and approaches is made. 
Annual reports and work programmes should be prepared for consultation with administration 
officers and review by colleagues or willing experts. 

Publication of results should be seen as an obligation to colleagues, as a means of 
informing sponsors and administrators of progress (positive and negative results) and 
problems, and as an intellectual discipline for the tree-breeding team. 


Planning an efficient tree improvement programme tailored to the requirements of 
forest management and utilization and resources available will involve the following main 
activities and decisions: 

(i) Define appropriate aims for the programme, especially concerning its duration, 
the improvements sought and the characteristics that are best manipulated by genetic means. 

(ii) Ascertain the resources that will be made available at least for the first 
several years. Appoint and encourage an interested tree breeder to help plan and run the 
programme with appropriate assistants. 

(iii) Locate suitable populations, in terms of provenance, size, age, sites, etc., 
for selection. 

(iv) Assemble relevant information on the biology of the species and decide the 
areas in which further research is required. Confirm or develop suitable techniques for 
grafting, pollinating, raising plants, etc. 

(v) Choose appropriate selection criteria, and limit their number, 
(vi) Develop methods for rapidly obtaining seeds of better quality. 

- 205 - 

(vii) Work out long-term dynamic, flexible breeding strategies no that a 
sufficiently large number of trees can be established in the breeding population, taking 
into consideration the planned programme and the number of generations for which 
improvement programmes are planned. 

(viii) Work out dynamic long-term and flexible breeding strategies having regard 
to aims, resources, time-scale, seed requirements, species characteristics, other 
institutes, regions and countries working on the same species should be carefully 

(ix) Set up efficient administrative arrangements and make provision for periodic 
reviews of the programme and its modifications as necessary. 


Allard, R.W., 1960. Principles of Plant Breeding. (Wiley, N.Y.). 

* An6n, (1978). Documentos de la Tercera Consulta Mundial Sobre Mejora de Arboles Fores tales. 

CSIRO. Canberra, Australia. 

Brown, A.G.. 1973. Notes on IUFRO Quantitative Genetics Workshop - Tokyo, 1972. Proc. 
Third Meeting Reps., Rec. Working Group No. 1, Mt. Gambler, 1972. App. 17: 1-6. Forestry 
and Timber Bureau, Canberra, Australia. 

Burdon, R.D., and Shelbourne, C.J.A., 1971. Breeding Populations for Recurrent Selection - 
Dilemmas and Possible Solutions. N.Z. J. For. Sci. 1 (2): 174-193. 

* Burdon, R.D. and Shelbourne, C.J.A., and Wilcox, M.D. Advance Selection Strategies. Proc 

3rd World Consult. For. Tree Breeding. FO: FTB- 77-6/2. Canberra, Australia. 

* Burley, J., and Nikles, D.G. (EdsJ Selection and Breeding to Improve Some Tropical Conifers. 

I, II. Commonw, For. Inst. Oxford, U.K. (1972/73). 

Burley, J., and Nikles, D.G., 1973. (Eds.) Tropical Provenance and Progeny Research and 
International Cooperation. Commonw. For. Inst. Oxford, U.K. 

Burrows, P.M., 1967. Seed Orchard Systems for Tree Breeding. Rhod. Zamb. Mai. J. Agrio. 
Res. 5: 273-280. 

Burrows, P.M., 1970. Co ancestry control in forest tree breeding plans. 2nd meeting of 
Working Group on Quant. Gen. Eds. Namkoong and Stern. South. For. Exp. Sta. New Orleans. 

Co sco, J.N., 1970. Genetic Selection Criteria in Pinus radiata. M.Sc. Thesis'. 122 pp. 
Aust. Nat. Univ., Canberra, Australia. 

porroan, K.W. , 1976. The Genetics and Breeding of Southern Pines. USDA FS Handbook 471. 

Eldridge, K.G., Brown, A.G. and Matheson, A.C., 1977. Genetic gain from a Pinus radiata wood 

* FAO, 1973. Informaci6n sobre Recursos Genticos Forestales. Doc. Occasional For. 1973/1. 

38 pp. (FAO. Rome) 

* Faulkner, R. 1975. Seed orchards. Gt. Br. For. Coram. Bull. 54 

* Kemp, R.H., BurUy, J., Keiding, H., and Nikles, D.G., 1972. International cooperation in 

the exploration, conservation and development of tropical and sub-tropical forest gene 
resources. Seventh World Forestry Congress. Argentina. 1972. 

Libby, W.J. f 1969. Seedlings vs. vegetative orchards. FAO - North Carolina State Univ. Forest 
Tree Improvement Centre, 1969. Lecture Notes: 306-16. School of Forest Resources, N.C. 
State Univ., Raleigh. 

* Libby, W,J. f 1973. Domestication strategies for forest trees. Can. J. For. Res. 3, 265-76. 

- 206 - 

* Namkoong, G., 1966. Foundations of Quantitative Forest Genetics* 85 pp. The Govt. For. 

Expt. Station of Japan. 

Namkoong, G., Snyder, E.B., and Stonecypher, R.W. Heritability and Gain Concepts for 
Evaluation Breeding Systems such as Seedling Orchards. Silvae Genetica 19(3): 76-84. 

Nikles, D.G., 1970. Breeding for Growth and Yield. In; Forest Tree Breeding. Unasylva 24 
(2-3), 97-98: 9-22. 

Nikles, D.G., 1973-a. A Proposed Breeding Plan for Improvement of Caribben Pine (Pinus 
caribbaea More let var. hondurensis Barr. and Golf.) Based on International Cooperation. 
In: Burley, J., and Nikles, D.G. (Eds.), 1973 - Selection and Breeding to improve Some 
Tropical Conifers. 2. (Commonw. For. Instit., Oxford, England). 

* Nikles, D.G., (1973-b), Biology and Genetic Improvement of Araucaria cunninghamii Alt. 

in Queensland, Australia. In; Burley, J., and Nikles, D.G. (Edsj, 1973 - Selection and 
Breeding to Improve Some Tropical Conifers. 2. (Commonw. For. Instit., Oxford, England). 

* Nikles, D.G., 1974. Planning a tree improvement program. Report on the FAO/DANIDA Training 

Course on Forest Tree Improvement. FAO/DEN/TF 7127 (FAO, Rome) 22 5-4 2. 

* Nikles, D.G., Burley, J., and Barnes, R.D., 1978. Progress and Problems of Genetic Improvement 

of Tropical Forest Trees. I, II. Commonw. For. Inst., Oxford, U.K. 

* Porterfield, R.L., 1978. Economic Evaluation of Tree Improvement Programmes. Proc. 3rd 

World Consult. For. Tree Breeding. FO: FTB-77-5/2. Canberra. 

* Reilly, J.J., and Nikles, D.G., 1978. Benefits and Costs of Tree Improvement; Pinus 

caribaea. Proc. 3rd. World Consult. For. Tree Breeding. FO: FTB-77-5/3. Canberra. 

Roche, L. (Ed.),* 1978. Metodologia de la Conservaci6n de los Recursos Gene*ticos Forestales. 
FO: MISC/7518. FAO. Rome. ^ 

Shelbourne, C.J.A., 1969. Tree Breeding Methods. Tech. Paper No. 55. 43 pp. New Zealand For. 
Serv. , For. Res. Instit., Rotarua. 

* Shelbourne, C.J.A., 1971. Planning Breeding Programme for Tropical Conifers. N.S. For Serv. 

Reprint No. 548. 20 pp. For. Res. Instit., Rotarua. 

Stonecypher, R.W. , 1966. The Loblolly Pine Heritability Study. Internt. Paper. Co. Tech. 
Bull. 5:1-128 

Stonecypher, R.W. , 1970. Multiple trait breeding. Unasylva 24 (2/3), 48-51. 

* Teich, A.M., and Carlisle, A., 1978. Anllisis de Costos y Beneficios en Prograroa de Mejora 

Genetica Forestal. Unasylva Vol. 30, No. 118/119. 

* van Buijtenen, J.P., 1975. The planning and strategy of seed orchard programs, including 

economics. In: Faulkner, R. Seed Orchards, Gt. Br. For. Comm. Bull. 54 

van der Meiden, H.A., 1978. Economics of Poplar Breeding. 3rd World Consult. For. Tree 
Breeding. FO: FTB-77-5/4. Canberra, 

Willan, R.L., and Palmberg, C., 1974. Improved use of forest genetic resources. Report on 
the FAO/DANIDA Training Course on Forest Tree Improvement. FAO/DEN/TF 112 (FAO, Rome). 

Wright, J.W., 1976. Introduction to Forest Genetics. (Academic Press, N.Y.). 

Yeatman, C.W., 1973. Gene Conservation in Relation to Forestry Practice. Thirteenth 
Meeting of the Committee on Forest Tree Breeding in Canada. Proc. 2: 13-17. Canadian 
For. Serv., Ottawa. 

* Zobel, B.J., 1973. Gene Preservation by Means of a Tree Improvement Programme. Thirteenth 

Meeting of the Committee on Forest Tree Breeding in Canada, 1971. Proc. 2: 13-17. 
Canadian For. Serv., Ottawa. 

Zobel, B.J. and McElwee, R.L., 1964. Seed Orchards for the Product ion of Genetically Improved 
Seed. Silvae Genetica 13 (1-2); 4-lTT 

Zobel, B.J., Weir, R.J, and Jett, J.B., 1972. Breeding Methods to Produce Progeny for 
Advanced-generation Selection and to Evaluate Parent Trees. Can. J. For. Res. 2; 339-345 . 

The references nmked with an asterisk (*) have already been quoted in the tt of the report. 

- 207 - 










4 e 



b ^ 6 + 



g ^ 






1 s 




s I 

O H t 


O C 4* C H J* 


' * O 4 H * C 






* i 












Firtt f 

cloaal ore 



- 208 - 

























-< O 












S -H 

3 00 



u-i cd 
O 1-1 

00 x: 

C 00 > 
^ C J= 

c -^ o 

3 .* <3 yj 

- M - cd n 7 

E- -O 

5 ^ 

O en 

~j w 

O H 


<- U- 

o- en 

W V2 

z o 
o o z 

Q ^3 h-4 

85 fc 

CUi c-i 

Cx3 fa tD 
2 O O 




" M 












O . 













































r \ 







etf O 


S 2 

O M 


o M 




Q ^ 

en z 


J- C 

>% * 

a < 

tj o2 

y^^ tjj 



r^^ fj 






w o 

h Qj 


^J en 

l-t 0> 

a- en 

O Q 



- 210 - 

Figure 4. Stages in the use o Eucalyptus cuttings 



Stump coppicing 


Collection of shoot crop 
lut tings placed in mist conditions 


13 F* 



NO. 2 


The sapling will have revealed 
its potential between 4 and 10 
years of age 

Well exposed stumps cut in November 

3 to 6 rotations on the stump during 
the same year 


Clonal test: 50 saplings spaced at 

4 x 4 m 

In the clone bank they are spaced 

at 5 x 5 m 


Coppicing of stumps 
at three years old 


Coppicing starts in October 

Gathering of shoots 


Cuttings under mist 

_ %l* 


2 rotations on the stump during 
one year 




1st coppicing at 3 years 

r The second generation clonal 
bank is created depending on 
results from clonal trial No. 1 
It receives the same treatment 
as the first generation clone 
bank. One third is coppiced 
each year 

Coppice rotation 3 years 




- 211 - 










u z 


C co 






H (d 






E u 

"~ UJ 

u. ? 
O H 

M >- 




(O W 













9 2 

et o 

0Q H4 

= 1 

W M 









co Id 


H U 









- 212 - 







- 213 - 


Marcelino Quljada R. 

Institute de Silvicultura 
Universidad de Los Andes 
Mfirida, Venezuela 


In Venezuela, as in many Latin American countries, hardwood species have not played 
a very important role in plans for genetic improvement, most of which are only in their 
initial stages. This is due largely to the fact that the natural forests, which are rich 
in flora, particularly broadleaved trees, have so far been expected to meet only local 
requirements, one species being replaced by another as time goes on. Another factor is 
the relatively low demand, particularly as regards quality, in many tropical regions. 
The weakness of the forestry sector within national economies has also had an effect: 
the forest enterprises are usually small, and though they use forest resources, they do 
not generate them. The larger enterprises have been centred around the coniferous 
species, particularly those belonging to the genus Pinus , for pulp and paper making. 

The history of the native broadleaved species can therefore be summed up as 

a) Deterioration of the genetic base of species which have been of commercial 
importance at one time or another. Typical examples are the species of the Mellaceae 
family. This situation is due basically to continued dysgenic selection in the natural 
range of these species. This kind of negative selection has well been called trader's 
selection since the best trees in the forest have been logged, leaving as a source of 
regeneration the trees with the worst phenotypic characteristics. 

b) Ignorance of the inherent and cultural properties of the species. This is 
due in large part to the feeling that the natural forests are inexhaustible, and to the 
problems in the initial utilization of some species. The latter is influenced by the 
use of genetically poor basic material and the application of cultural techniques 
developed specifically for temperate-zone species. However, the situation has begun to 
change in many places, due to the following factors among others: 

1) exhaustion of natural sources of wood, particularly near large population 
centres ; 

2) an increase in demand owing to the growth in population; 

3) the need to diversify economies traditionally dependent on a few 
resources, particularly non-renewable resources. Here forestry 
economics is particularly important ; 

4) the emergence of a desire to recuperate natural ecosystems and create 
recreation areas. 

Some countries have proposed as a solution the use of introduced broad leaved 
species, such as Tectona. Gmelina and Eucalyptus, in view of their favourable produc- 
tivity characteristics. 

- 214 - 

Even so, it has become necessary to increase the use of native species, either as 
means of recuperating and preserving the autochthonous ecosystems, or because of the fear 
that after a time the introduced species may present problems which outweigh the advan- 
tages they offer. The main advantage of the native species, in any case, is biological 
adaptation they have achieved in certain sites through generations of natural selection. 


If greater use is to be made of native species in genetic Improvement programmes, 
it is essential that a minimum amount of knowledge on these species be available, covering: 

a) Natural variation in those areas of the country where the species of interest 
are still found, in order to know the morphological, anatomical, physiological and techno- 
logical variants that imply hereditary adaptation to various site conditions for selection 

b) Phenology of the species, in order to assess the seed production capacity for 
supplying plantation programmes. Flowering habits also need to be known for the purpose 
of controlled pollination. 

c) Cultural habits, from the nursery production phase to plantation techniques and 
methods and subsequent tending. The purpose is to reduce the Influence of this component 
as much as possible, maximizing the genetic potential for use of plantation sites. Of 
particular interest in this phase, for the purposes of Improvement, is the species vege- 
tative propagation habit. 

d) Development habits under cultivation, including quantitative and qualitative 
growth and technological properties. This knowledge serves as a basis for estimating 
gains in the possible selection systems that may be applied. 

Once this knowledge is available, genetic improvement would be directed towards 
aspects of a general nature and of a particular nature. The first category, common to all 
species, includes the growth phase, which is essential for a greater yield per unit of 
area and for enabling native species to compete with Introduced ones - an aspect which 
is very commonly emphasized in justifying the introduction of species. In this connec- 
tion the programmes would be directed towards the formation of quick-growing subpopula- 
tions, through the selection and testing of vegetative propagules or seed progeny. 

In the second category account would be taken of particular problems of species or 
groups of species which might result, through appropriate crossing, in a more effective 
use of the product. In this connection special mention should be made of the following, 
which are among Venezuela's most valuable timber species. 

a) Bombacopsis quinata. Efforts are being made to control the development of 
buttresses, in order to increase the amount of utlllzable bole wood for sawmill Ing. 
Since the buttresses, like the thorns, seem to be a character inherently linked with the 
evolution of the species, the aim is not to eliminate them completely, but to reduce 
their development to a minimum, through a process of selection and crossing specific 
individuals, on the basis of observations and preliminary results with forest trees and 

b) Meliaceae species. Of particular interest are species of the genera Cedrela and 
Swietenia, because of their resistance to Hypsipyla. Since in most part of the world 
where these species grow naturally they have been subjected to intensive dysgenlc selec- 
tion, the genetic base seems to be fairly weak. A possible improvement strategy would be 
to effect intra-specles selection followed by inter-species hybridization, with possible 
subsequent backcrossing. Crosses between Cedrela odorata, . angustifolia, . fissllis 
an< * PQPtana deserve particular attention, and knowledge of the floral biology of these 
species should be intensified. Crosses between Swietenia macrophylla, J5. humills and 

Are also being considered. Spontaneous inter-specific hybridization in this 

genus has already been reported. The possibility of Intergeneric hybrids, e.g. Cedrela x 
Toona, is not to be rejected. The problems Involved may be considerable, but the genes 
are morphologically so similar that until fairly recently Toona cillata was classified 
under the genus Cedrela. 

- 215 - 

c) Tabebuia rosea. Efforts will be made to correct ramification habits, particu- 
larly the typically polydichotomlc bifurcation of the main trunk. This phenomenon 
appears to be connected with Inadequate individual sources of seed and high herltablllty 
of the characteristic. Strict individual selection, particularly since stands with very 
good Individuals are now available, should lead to important achievements in the near ' 

d) Phitecelobium saman. The attractive feature of this species is its trunk- 
crown ratio. In the open, under a sllvo-pastoral system, the tendency is towards the 
development of a very broad crown, to the detriment of the bole quality. This tendency 
has also been observed in the natural forest, where, however, the trunks are better. 
Through a process of continuous individual selection, with interprovenance crossings, 

a broad crown with better trunk quality can be sought in the open field, and a smaller 
crown and thicker trunk in dense plantations. 

- 216 - 





Marcelino Quljada R. 

Institute de Silviculture 
Universidad de Los Andes 
Mfirida, Venezuela 


Bombacopsis quinata is one of the most important dry tropical forest timber 
species in Venezuela. It is found on various kinds of soils ranging from well drained 
terrace soils to poorly draining swamp soils. 

Rainfall distribution In the Bombacopsis quinata range oscillates between 1 000 
and 2 000 mm/yr, with a 3-4 month dry season from December to March. The mean annual 
temperature is roughly 27C, with maxima of 32C and minima of 22C. !B. quinata 
wood is widely used in construction, interiors, exteriors, furniture, cabinet making, 
carpentry, etc. The wood density is roughly 0.40, based in kiln dry weight and 
volume in green wood. 

. quinata has been planted on a moderate scale to enrich natural forest, parti- 
cularly in the forest reserves. Various trials have Indicated open field planting as a 
promising technique, assuming proper site selection. 

Studies on I*, quinata were begun in 1961 to gain a better understanding of this 
species for breeding purposes. Initially, vegetative propagation habits were studied. 
Bombacopsis quinata responded so well to these trials that it came to be considered an 
easily propagated species. Other studies on the flowering and fruit setting habits 
were begun in 1968 with the establishment of seed orchards, including open and guided 


Bombacopsis quinata are monoic, with hermaphrodite, generally erect flowers. The 
stamens surround a pistil with stigma slightly protruding beyond the anthers. 

The flowers, 10-15 cm long, are night-blooming. They usually open after 6 p.m., 
mostly from 8-10 p.m. , when the temperature drops below 25C and the relative humidity 
rises above 60 percent. The morning after the flowers open, they drop their floral covers 
and stamens. The style remains, even after the fruit has formed. The flowering season 
runs from the end of October to March and sometimes April, depending on how long the dry 
season lasts. The best flowering periods are January and February, which are usually the 
driest months. Observations in seed orchards and natural forests Indicate gteat varia- 
bility in flowering. Some trees have a prolonged flowering period which lasts throughout 
the season. Some flower early in January and some flower late, from February onwards. 
This is very important in planning the lay-out of seed orchards. The initial age of 
flowering, seams to indicate early flowering. Trees not yet 10 years old have been 
observed to flower. Extreme cases of plants grown from seedlings, flowering after only 
six months in the field, have also been observed. 

- 217 - 

Fruit setting begins in January, with a lapse of 50-60 days from flower opening 
and pollination to dehiscence. 

The fruit is dehiscent. The seeds are fuzz-covered and therefore easily dispersed 
by wind. 

Seeds from open pollination in seed orchards average 47 per fruit, ranging from 
zero (empty pods) to 140. The average in the natural forest is 30 seeds per fruit, with 
heavy insect attacks on fruit and seeds. Viable seeds are usually smooth and fairly 
resistant when pressed flat with the hand. Non-viable seeds are usually rough and are 
easily flattened with a slight pressure of the fingers. 

Colouring ranges from light to dark brown with various degrees of segregation 
within the raroets of a clone, which seems to be linked to the source of pollen. 

Size is heavily influenced by the mother tree, and is reflected in the number 
of seeds /kg: 20 000 to 40 000, averaging 32 000 seeds /kg. 

Studies Indicate that the species is highly resistant to self ing, which favours 
cross-pollination. The low occurrence of selfing seems to be facilitated by the fact 
that as the flower is hermaphrodite, the stigma is receptive at the moment the flower 
opens but before it scatters its pollen. Pollen dispersal usually takes place 10-15 
minutes after the pollen is exposed to wind. Some system of reproductive incompatibility, 
likewise favouring cross-pollination is also conceivable. 

The findings of several years of seed germination trials on seeds from over 

100 seed orchards and natural forest trees, suggest that _B. quinata can be classed as 

a genetically clean species inasmuch as negative traits such as albinism or dwarfism 
have not appeared to date. 


The size and open floral structure of the J3. quinata flower has made it much 
easier to artificially pollinate the species. 

When the flower reaches normal size, both the bud and the flower are isolated in 
a cloth sack to avoid damage from insects looking for nectar in the flowers. For arti- 
ficial pollination, breeders allow the flower to open naturally. When it opens, the 
stamen is cut away with small scissors, carefully avoiding contact with the stigma of 
the pistil. The breeder takes advantage of the fact that at this stage, the pollen is 
still not easily dispersed. 

Once the plant has been emasculated, the stigma is introduced into the appropriate 
pollen source, and completely saturated. The cross is then identified and the pollinated 
flower is again covered with a cloth bag, perhaps until the fruit falls off but in any 
case until such time as the fruit is clearly formed. 

The pollen is collected at emasculation. The stamens are placed in a Petrl dish 
and allowed to dry for at least one hour after which the breeder shakes the dish to 
release the pollen* The pollen is stored in the same container. It has been stored 
in an ordinary refrigerator for up to two weeks, with no great loss of viability, as 
long as the Petri dish was kept tightly closed and excessive accumulation of moisture 
within the container avoided. Controlled pollination has been 50 percent successful, 
with figures of up to 60 percent for cross-pollination and 4 percent for selfing. 
Variations have been observed in inter-crossing different clones, with the highest per- 
centage of success from inter-crossing clones of different provenances. 

- 218 - 

Success in cross-pollination has varied greatly: very poor years with rates as 
low as 25 percent and very good years with rates as high as 80 percent, apparently 
related to environmental determinants affecting floral development. 

Seed production under controlled pollination can go as high as 87 seeds per fruit, 
which is 1.9 times the yield from open pollination in seed orchards. Here also there 
have been extreme variations, ranging from empty pods to 150 seeds per fruit. 

No great qualitative differences have been noted in the viability of selected seed 
with either method. Germination rates have averaged about 90 percent. 

Seed has been maintained in the laboratory in paper envelopes for as much as six 
months and remained highly viable. At temperatures of around 5C, the seed can remain 
under ground in a closed glass container for more than a year without any visible signs 
of deterioration. Seed has been stored for about two years in a 40-80 percent sulphuric 
acid environment in a chemical drier, the only precaution being to check that the paper 
or cloth packages in which the seed was stored had not been corroded by the acid, and 
to change them every six months. 

- 219 - 



Climatic feature I/ 

Coordinates: 62W, 8<>N 

Altitude: 50 m above sea level 

Temperature: Mean max. 32oc 
Mean min. 2QOC 
Mean annual 26oc 

Rainfall: 1150-1200 mm/yr, with the main rainy season from June to September and 
the shorter rainy season from November to mid-January 

Winds: Average wind speed 7.8 km/hr, with occasional gales of 60-80 km/hr 
Evaporation: Mean annual values: 2000-2100 mm. 

Plantation Programme (Pinus caribaea var. hondurensis; Eucalyptus spp. <1%) 

Company Site 














- CONARE Coloradito 

60 000 

7 000 

5 000 


100 000 

22 000 

5 000 






12 000 

4 000 

10 000 

- CVG Uverito 

150 000 

50 000 

7 000 


322 805 

83 000 



- Forester 

60 000 

2 600 

10 000 

- Sipas 

30 000 

3 000 

1 500 

- Guayamure 

60 000 

7 000 

1 500 


150 000 

12 600 



4 000 

TOTAL Llanos Orientales 2/ 

472 805 

99 600 

I/ "Uverito - un bosque en la Sabana". Divisi6n de Relaciones Pfiblicas, Corporaci6n 

Venesolana de Guayana, CVG. Publicaci6n No. 1078-652; Tomo I, Folio 20; October 1978. 
I/ For country-wide information, see Appendix VI, "Venezuela 11 . 

- 220 - 



(P. caribaea var. hondurensis. Uverito) 


Human Resources 

Total Production 

Seedbeds ; 

Prepare terraces 




Prepare containers: 

Cut cardboard 

Staple containers together 

Fill, weigh 


Extract seedlings 
Move to terraces 
Open seed holes 
Sow in containers 

Field work: 

Sprinkler irrigation 
Fungacide treatments 


Transport containers 
Line up containers 
Fill containers with sand 
Compact sand 

1 foreman 
8 workmen 

1 supervisor 

1 foreman 

10 cutters 

60 helpers 

6 fillers 

4 foremen 
1 helper 
6 hoers 
32 sowers 
6 hole-diggers 

1 foreman 
14 workers 

360 000 plants /day 

3 foremen 

6 people to line up containers 
36 helpers 
27 people to fill containers 

9 people to pack sand 


Land preparation: 

Demarcation of stands 
Soil preparation 
Insect control 

4 supervisors 

8 foremen 

64 people to fill boxes 
90 planters 

20 people to spray insects 
24 tractor drivers 

7 000-10 000 ha/yr 

(area planned for 
1980 - 7 000 ha; 
5 000 ha in containers, 
2 000 ha directly 

- 221 - 

ANNEX V/3 (cont'd) 

Operation Human Resources Total Production 

Planting: (cont'd) 


'J *"* 

Fill boxes and trucks ao So 

Transport plants a a 


o o 

Maintenance ; *> *> 

^^ ^ <y QI 


Open up and weed firebreaks * 

Prevent and fight fires o> 

Control weeds ** 
Insect control 

Additional personnel, 
to staff dining rooms, 
infirmary , machine shop, 
carpentry shop, etc. 


- 222 - 


(Plnus carlbaea var. hondurensis. Uverlto) 








1 1 

1 1 

1 1 



1 I 



1 1 


j i 
i i 


i i 








! ! ! ! 



i i i 


: i i 




1 ! 



-^ 1 


^ i 




1 1 1 



T 1 i ! 










^ J^ 




co a 



Seeding in seedbed 
Prepare containers 
Prepare terraces 
Transplant seedlin 

Sprinkler irrigati 
Trimnlng roots 
Insect control 





2 3 


Fill crates and tr 
Actual planting 

Insect control 







CO > 






848 S 

H 4J CO 




9 "> S 



60 2 





> -H (3 











- 223 - 




1.1 Aera of country: 2 790 000 km 2 

1.2 Location: longitude: 53 - 73W; latitude: 21 - 55 S. 

1.3 Population: 26 000 000 

1.4 Main climatic and vegetative zones: "Selva Misionera" (Misiones Forest) - "Selva 
Tucumano Boliviana" (Bolivian Tucuman Forest) - "Selvas en Galerla" (Riparian 
Forests) - "Bosques Subantarticos"(Subantartic Forests) - "Parque Chaqueno" (Chaco 
Park) - "Parque Puntano-Pampeano" (Puntano Pampas Park) - "Parque Mesopotamico" 
(Mesopotamian Park) - "Estepa Pampeana" (Pampas Steppe) - Estepa Patagonica" (Pata- 
gonlan Steppe) - "Monte Occidental" (Western Forest) and "Estepa Punena" (Puno 
Steppe) (see Annex I). 


2.1 Area of forests: 60 300 000 ha (productive forests: 39 000 000 ha), 

2.2 Proportion of land under forest: 20.54%. 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

a) Rational utilization in accordance with forestry plans. 

b) Utilization with special emphasis on the most valuable end product. 

c) Encouragement of reforestation and improvement of national forests. 

d) Encouragement of afforestation, encouraged by development credit. 

e) Conservation of protective forests. 

2.4 Legislation available to implement policy? 
YES - Law No. 13.273 of 1948 

2.5 Ownership of forests (out of the 39 000 000 ha of productive forests) 

Under State control: 18 150 000 ha (46,16%) 
Private ownership: 20 850 000 ha (53.84%) 
Community ownership: 
No effective control: 

2.6 Major forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) : 

In 1977: _ _. ,.__ 

Saw-wood 21 000 000 m 2 Plywood 54 000 t 

Veneer logs 58 000 m* Charcoal 278 000 t 

Logs for fibreboard 225 000 m3 Fuelwood 1 095 000 t 

Logs for sleepers 61 000 t Pulp for paper 703 000 t 

p s t8 199 000 t Tanin (extract) 87 000 t 

2.7 Staff 

(with diploma or 
certificate of training) 

2.8 Gross annual budget for forestry: US$ 47 680 000 21. 

I/ Only under the National State. . 

|/ Only under the National State, including incentive subsidies (US$ 1.00 - $a 1.630). 

- 224 - 






Net total area of plantations at the end of 1978: 530 

000 ha - 1 . 


Planned annual target of afforestation/reforestation: 90 

000 ha/year (1979) -' 


Organization and administration of planting schemes: 


State forest services: 100% 


Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 

pulpwood, fuelwood, protection, etc.). 


Indigenous species 

Mean annual increment 

(without bark) at the 

Net area Rotation 

end of the rotation 

Product/Purpose Species (ha) i/ (years) 

(m 3 /ha/year) 

Saw-wood Araucarla -/ 

17 (man-made planta- 

Pulpwood augustifolia 30 000^' 40 


Tanin and Schinopsis 

6 for the stand, 0.5 

sleepers sp. 12 000 000 120 

for the species 

Saw-wood Nothofagus 


sp. 1 000 000 90 

(homogeneous forest) 

Saw-wood ( Cedrela sp. ) 


( Cordia sp. ) 2 500 000 60 

0.15 (for the stand) 

( Half our odendron 

. 3 (heterogeneous 

( sp. ) 



Introduced species 

Mean annual* Increment 

(without bark) at the 

Net area Rotation 

end of the rotation 

Product /Purpose Species (ha) i/ (years) 

(m 3 /ha /year) 

Paper pulp Pinus elliottii) ,- n nftn 10-25 


and resin Pinus taeda ) " u uuu 10-25 


Wood and pulp Pseudotsuga 

menziesil 32 


Pinus radlata 30 


Pinus ponderosa 35 


Containers pulp Populus sp. 170 000 10 


Eucalyptus sp. 100 000 20 





Service and control of seed supply 


Have seed zones been defined for Indigenous species (I.e. 

, is it common to 

restrict the use of seed for afforestation/reforestation 

to the particular 

zone in which it was collected)? 

YES (only with Araucaria angustifolia in Mlsiones) 


Is there a national seed certification system? NO 4/ 


Are there facilities for storing seed at controlled temperatures? YES 


Does the supply of seed cover the demand for the species 


In 3.3.1? NO in 3.3.2? NO 

T7 The net area is the gross area of plantations, less the area occupied by roads, paths, 

firebreaks, buildings and other areas not covered by trees. 
2J Approximate dates. 

3J Corresponds to range of the species in Argentina. 
\l In process; status: preliminary draft law. 

- 225 - 


5.1 Does the country have an official tree improvement programme? YES I/ 

5.1.1 If yes, list the species concerned: 

Pinus elliottii Populus sp. 

Pseudotsuga menziesii Salix sp. 

Pinus ponderosa Eucalyptus saligna/grandls 

Araucaria angustifolia Melia azedarach 

5.1.2 Most important characters to be improved or bred for: 

1) Productivity 

2) Ecological adaptability 

3) Disease resistance 

4) Seed quality 

5) Wood quality 

5.2 Brief outline of improvement methods already applied. 

5.2.1 Species /provenance trials (indicate in parentheses the number of provenances 
being tested): 

a) Pinus elliottii and P. taeda (38 and 29 provenances) 27 

b) Eucalyptus camaldulensis, E. viminalis (33 and 22 provenances) 2/ 

c) Pinus in northwest Argentina (19 species) 2/ 

d) Pinus on the Argentine coast (17 species) 2J 

5.2.2 Area of seed stands in each of the main species: _3/ 

Pinus ponderosa 8 ha 

Pseudotsuga mensiesii 15 ha 
Araucaria angustifolia 414.5 ha 

5.2.3 Plus trees of the main species (indicate in parentheses the number of trees): 3f 

Pinus ponderosa (24) 
Pseudotsuga menziesii (17) 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 31 NO 

5.2.5 Progeny testing? (species and area): 31 

Pinus ponderosa (1979) 
Pseudotsuga menziesii (1979) 

5.2.6 Other methods of improvement (specify): J3/ NO 

Species Method % of success 

Pinus elliottii Stakes (horm. treatment and Started on 12.12.79 

treatment of mother tree) 

Eucalyptus saligna " Started on 12.12.79 

Pseudotsuga menziesii " Starts on 1.2.80 

Pinus ponderosa " Starts on 1.2.80 


Plfease list the references to tree improvement in the country, in publications, 
reports, etc. 

IDIA (Suplemento Forestal No. 6)- INTA - Bs.As. 1970 

IDIA (Suplemento Forestal No. 8)- INTA - Bs.As. 1973 

IDIA (Suplemento Forestal No. 5)- INTA - Bs.As. 1966 

IDIA (Revista No. 12) - Buenos Aires 

T/ By 1980 at national level. 

2/ Institute Nacional de Tecnologla Agropecuaria (INTA) - Natl. Institute of Agricultural 

31 Only under the jurisdiction of the National Forestry Institute. 

- 226 - 


Phytogeographical regions 


Sub-in! rrrir fart. northern 

Sub-ant *rrtir for*t 4.i<HtHern 

Argtnttn* Antiirctir f 

South Ortn+r ^ u 

- 227 - 



1.1 Area of country: 1 058 000 km 2 

1.2 Location: longitude: 5726 f - 6038'W; latitude: 938 f - 22O53 f N 

1.3 Population: 4 647 816 (September 1976) 

1.4 Main climatic and vegetative zones: mountain-building in the area has produced 
four different zones: warm valleys, valleys, high plateaux and plains. 


2.1 Area of forests: 56 468 000 ha 

2.2 Proportion of land under forest: 51% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

a) To formulate the country's forest policy and its plans for implementation. 

b) To manage national forest resources on a permanent basis. 

c) To promote and/or make a forest Inventory for Bolivia. 

d) To authorize, guide and supervise forest utilization, in accordance with 
the provisions of the law. 

2.4 Legislation available to Implement policy? YES 

2.5 Ownership of forests: 

Under State control: 99.9 percent 
Community ownership: 0.1 percent 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) : 

Veneer logs, sawn-logs, sleepers, posts, planting for mines, building timber, 
charcoal fuelwood, rubber and chestnut. 

2.7 Staff State Other 

Professional 28 5 

Sub-professional (with diploma or certificate of training) 21 5 

2.8 Gross annual budget for forestry: US$ 1 000 000. 

3*1 Areas 

3.1.1 Net total area JL/ of plantations at the end of 1978: 10 000 ha. 

3.1.2 Planned annual target of reforestation: 1 500 ha/year. 

3.2 Organization and administration of planting schemes: 

3.2.1 State Forest Services: 90 percent 
C^MARAFOR: 3 percent 

Other private owners: 7 percent 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and 
stakes, pulpwood, fuelwood, protection, etc.). 

I/ The net area Is the gross area of plantations, less the area occupied by roads, 
paths, firebreaks, buildings and other areas not covered by trees. 

. 228 - 

3.3.1 Indigenous species 

Product /Purpose 



Net area 

(ha) !/ 


Mean annual Increment 
(without bark) at the 
end of the rotation 











3.3.2 Introduced species 

Product /Purpose 






Eucalyptus spp 


Met area 
(ha) I/ 

9 000 




Mean annual increment 
(without bark) at the 
end of the rotation 
(m3 /ha/year) 



4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which it 
was collected)? YES 

4.1.2 Is there a national seed certification system? NO 

4.1.3 Are there facilities for storing seed at controlled temperatures? NO 

4.1.4 Does the supply of seed cover the demand for the species listed 
in 3.3.1? NO 

in 3.3.2? NO 


5.1 Does the country have an official tree improvement progr ? 

5.2.5 Progeny or testing? (species and area): 

Eucalyptus rostrata Pinus radlata 

E. vimlnalls P. elllottli 

E. saligna P. taeda 

E. Globulus Cupressus lusitanica 

E. citriodora 

E. teretlcornis 

- 229 - 



1.1 Area of country: 8 511 970 km 2 

1.2 Location: longitude: 34S-74W; latitude: 5N-33S 

1.3 Population: 119 670 000 

1.4 Main climatic and vegetative zones: See Annex I 


2.1 Area of forests: 352 000 000 ha 

2.2 Proportion of land under forest: 41% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

a) Forest production to meet the National Programmes for pulp and paper and 
for the charcoal fueled steel industry 

b) Reforestation Programme for small and medium-sized rural holdings 

c) Settlement and rational utilization of the forest resources of the 
Brazilian Amazon region 

2.4 Legislation available to Implement policy? YES 

2.5 Ownership of forests 

Under State control: 75% 

Private ownership: 15% 
Community ownership: 

No effective control: 10% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) : 

Saw-wood 16 325 000 m 3 Charcoal 17 618 000 t 

Sawn wood 7 400 000 m3 Paper 2 045 969 t 

Pulp 1 253 784 t Veneer 695 000 m 3 

Fibre board 574 000 m 3 Particle board 461 000 m3 

Plywood 659 000 m 3 Fuelwood 140 000 000 m 3 

2.7 Staff State Others 
Professional 1 200 

2.8 Gross annual budget for forestry: US$ 30 266 343.83 


3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1978: 3 319 034 ha 

3.1.2 Planned annual target reforestation: 335 442 ha/year 

3.2 Organization and administration of planting schemes 

3.2.1 State forest services: 5% 

Others: 95% (most of this due to tax incentives) 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and 
stakes, pulpwood, fuelwood, protection, etc.). 

I/ The net area Is the gross area of plantations, less the area occupied by roads, 
paths, firebreaks, buildings and other areas not covered by trees. 

- 230 - 

3.3.1 Indigenous species 

Product/Purpose Species 



3.3.2 Introduced species 

Product /Purpose Species 

Pinus spp. 

Net area Rotation 
(ha) II (years) 

73 074.91 

39 744.75 



Mean annual increment 
(without bark) at the 
end of the rotation 


Net area 
(ha) !/ 


Mean annual increment 
(without bark) at the 
end of the rotation 

.05 533.82 
'39 781.35 




4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e., is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 

it was collected)? YES 

4.1.2 Do special seed production areas exist? YES 

4.1.3 Are there facilities for storing seed at controlled temperatures? YES 

4.1.4 Does the supply of seed cover the demand for the species listed 
in 3.3.1? YES 

in 3.3.2? NO 


5.1 Does the country have an official tree improvement programme? YES 

5.1.1 If yes, list species involved: 

Pinus caribaea var. hondurensis 
P. caribaea var. caribaea 
P. elliottii var. elliottii 
Araucaria angustifolia P. oocarpa 

5.1.2 Most important characters to be improved or bred for: 

Eucalyptus grandis 
E. urophylla 
E. saligna 

Volume (dbh and height) 

Wood quality 
Resistance to diseases 

5.2 Brief outline of improvement methods already applied: 

5.2.1 Species /provenance trials (indicate in parentheses the number of provenances 

being tested): 

Eucalyptus spp. 

Pinus caribaea var. hondurensis 
P. caribaea var. caribaea 
P. caribaea var. bahamensis 
P. oocarpa 

43 species, 1-37 provenances of each species 
(7) I/ 
(2) I/ 
(2) I/ 
(19) I/ 


Provenances introduced by PRODEPEF between 1971 and 1977. 

- 231 - 

5.2.2 Seed stands: area of seed stands In each of the main species: 

Eucalyptus grandis 114.99 ha 

E. urophylla 36.46 ha 

E. saligna 2.30 ha 

E, citriodora 8.15 ha 

E. paniculata 10.45 ha 

E, vimlnalis 215.28 ha 

Pinus oocarpa 1 069.58 ha 

P. carlbaea var. hondurensis 1 202.96 ha 
P. elllottli var. elliottii 30.90 ha 

P. taeda 106.50 ha 

P. carlbaea var. carlbaea 346.16 ha 

P. caribaea var. bahamensls 185.48 ha 
P. kesiya 117.87 ha 

P. strobus var. chap ens is 6.98 ha 

5.2.3 Plus trees of main species (Indicate in brackets the number of trees): 

Eucalyptus grandis (200) P. caribaea var. hondurensis (300) 

E. saligna (50) P. carlbaea var. bahamensis (300) 

Pinus oocarpa (100) P. caribaea var. caribaea (300) 

P. taeda (50) 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 

Eucalyptus grandis (100 clones) 
E. saligna (50 clones) 
Pinus caribaea (300 clones) 

5.2.5 Progeny testing? 

Eucalyptus grandis (15 tests) P. caribaea var. hondurensis (22 tests) 
E. saligna (5 tests) P. kesiya (10 tests) 

5.2.6 Other methods of improvement (specify): 

Area of seed collection 
Grafted seed orchard 
Obtaining and use of hybrids 


Species Method % of success 

Eucalyptus camaldulensis Estaqula 60 

E. tereticornis Estaquia 50 

E. grandis Enxertia 60 

E. dunnil Enxertia 70 

E. dunnii Estaqula 50 

P. kesiya Enxertia 80 

P. caribaea Enxertia 70 


Please list the references to tree Improvement in the country in publications, 
reports, etc. 

IPEF Revista IUFRO Proceedings 

SIF Boletin Tgcnico Revista Arvore 

IPEF Circular Tficnica PRODEPEF Sgrle Tgcnica 
Silvlcultura en Sao Paulo 


- 232 - 



Hqiics iiociiyiTiui tiw. * * *** 

- 233 - 

Country statement: CHILE 

22 2 

1.1 Area of country: 2 007 000 km - 757 000 km mainland and islands, and 1 250 000 km 


Area of coi 

1.2 Location: longitude: 6730' - 7530'W; latitude: 17 - 56S (mainland). However, 
it stretches down to the Antarctic. 

1.3 Population: 11 000 000 

1.4 Main climatic and vegetative zones: 

Climatic a) Sub-tropical arid 

b) Warm temperate with adequate humidity 

c) Rainy temperate 

d) Tundra 

e) Cold steppe 

f) Freezing climates 

Vegetative a) Xeromorphic zone 

b) Mesomorphic zone 

c) Hygromorphic zone 

d) Andean zone 


2.1 Area of forests: 20 000 000 ha 

2.2 Proportion of land under forest: 26 percent - taking into account both mainland 
and island Chile. 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? . 

To achieve accelerated growth and development of the forest sector through conserva- 
tion, protection, management and utilization of forest resources, bearing in mind 
the nation 1 s interests and the country's economic and ecological limitations. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: 

2.5.1 Under State Control: 60 percent 

2.5.2 Private ownership: 40 percent 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer 
logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Chemical pulp, mechanical pulp, newsprint, other papers, cardboard, fibreboard, 
particle board, plywood, veneers, logs, sawn timber, posts, stakes, sleepers, 
honey, bark, leaves, mushrooms. 

2.7 Staff State Others 

2.7.1 Professional: 60 300 

2.7.2 Subprofessional (with diploma or 

certificate of training): 100 360 

2.8 Gross annual budget for forestry: US$ 20 000 000. 

3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1978: 
* 712 000 ha 

3.1.2 Planned annual target/reforestation: 70 000 ha/year 

y The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 234 - 

3.2 Organization and administration of planting schemes 

3.2.1 State Forest Services: - 

3.2.2 Others (specify): 

Small private owners: 45 percent 
Large companies: 55 percent 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuel wood, protection, etc). 

3.3.1 Indigenous species: 

Mean annual increment 
(without bark) at the 

Product/ Net area Rotation end of the rotation 
Purpose Species (ha) I/ (years) (m /ha/year) 

Sawn wood Nothofagus spp. 500 annual 80 14 
2 000 total 

Forage Atriplex 400 annual 
repanda 2 000 total 

Bark Quillaja 100 annual 40 7 
saponaria BOO total 


Introduced species: 

Mean annual increment 
(without bark) at the 
Product/ Net area Rotation end of the rotation 
Purpose Species (ha) I/ (years) (m 3 /ha/year) 

Paper, sawn wood, Pinus 62 000 annual 
^>ulp radiata 640 000 total 20-25 22 

Firewood, posts, Eucalyptus 3 500 annual 
parquet globulus 40 000 total 15-20 25-30 

Matches, cases, Populus 2 200 annual 
sawn wood spp. 22 000 total 15-20 25-30 

Sawn wood Pseudotsuga 700 annual 
menziesii 2 000 total 80 13 

Forage Atriplex 600 annual 
nummularia 3 000 total 




Service and control of seed supply 


Have seed zones been defined for indigenous species (i.e. is it common to 
the use of seed for afforestation/reforestation to the particular zone in 
was collected)? For very few species. 

which it 


Is there a national seed certification system? YES 


Are there facilities for storing seed at controlled temperatures? YES 


Does the supply of seed cover the demand for the species listed: 
in 3.3.1? YES 
in 3.3.2? YES 




Does the country have an official tree improvement programme? YES 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 235 - 

5*1.1 If yes, list species involved: 

- Pinus radiata, through the establishment of seed orchards. 

- Various species involved in the provenance trials (see item 5.2.1). 

- Nothofagus alpina, an indigenous species, through a variability study. 
5.1.2 Most important characters to be improved or bred for: 

Growth rate Volume 

General form Wood density 

General adaptability Resistance to Dothistroma pini 

5.2 Brief outline of improvement methods already applied: 

5.2.1 Species/provenance trials (indicate in parentheses the number of provenances being 

Eucalyptus delegatensis (12) Pinus radiata (14) 

E. regnans (6) P. ponderosa (6) 

E. obliqua (4) P. contorta (3) 

E. bicostata (4) P. muricata (2) 

E. camaldulensis (3) Pseudotsuga menziesii (2) 

5.2.2 Area of seed stands in each of the main species: 
Pinus radiata: 120 ha 

5.2.3 Plus trees of main species (indicate in parentheses the number of trees): 
350 plus trees, with a selection intensity of 1 in 120 000. 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 

Only pinus radiata. There are two types: (1) clonal ; There are 8 of these, 

with 42 clones each and areas fluctuating between 11 and 30 hectares, 20 being 

the average; (2) from seedlings; there are 3 of these, with 200 families as a 
basis, 20 hectares each. 

5.2.5 Progeny testing? (species and area): 

In May 1980, 55 hectares on 14 different sites, will be put under progeny trials 
of Pinus radiata. 

5.2.6 Other methods of improvement (specify): 

38 000 seed trees of Pinus radiata have been selected, which have produced 7 000 kg 
of seed, with an expected improvement of roughly 5-6 percent. 


Species Method 7. of success 

Pinus radiata Lateral grafting 80 


Please list the references to tree improvement in the country, in publications, 
reports, etc. 

Burdon, R. Mejoramiento genetico forestal en Chile. Documento de trabajo No. 11. 
FO:DP/CHI/76/003 Investigacion y desarrollo forestal. Santiago de Chile, 1978. 

Delmastro, R. Proposicion de un programa cooperative de mejoramiento genetico entre 
la Universidad Austral, instituciones y empresas forestales. Facultad de Ingenierfa 
Forestal, Universidad Austral de Chile, Valdivia. 1976. 

Delmastro, R. Informe de genetica forestal. Proyecto CONAF/PNUD/FAO/CHI/ 76/003. 
Codigo III - 1.1 Facultad de Ingenierfa Forestal, Universidad Austral de Chile, 
. Valdivia, 1977. 

Moreno, D.G. Consideraciones preliminares para un programa de mejoramiento genitico. 
(To be published in May 1980 in the Supplement to "Chile Forestal"). 

Moreno, D.G. and Stutz, C. Informe de la gira a Nueva Zelandia sobre mejoramiento 
genetico. Documento interno presentado a la Direcci6n Ejecutiva de CONAF a la 
Direcci6n del Proyecto CONAF/PNUD/FAO/CHI/ 76/003. Chilian, Chile, 1979. 

- 216 - 

Delmastro, R. Primer inform anual. Convenio de mejoramiento genltico UACH - 
entpresas forestales. Facultad de Ingenieria Forestal, Universidad Austral de 
Chile, Valdivia, 1977. 

Delmastro, R. Segundo informe anual* Convenio de mejoramiento genetico UACH - 
empresas forestales. Facultad de Ingenieria Forestal, Universidad Austral de Chile/ 
Valdivia, 1978. 

Delmastro, R. Manual para los ensayos de progenie de polinizaci6n abierta. Convenio 
de mejoramiento genetico UACH - empresas forestales. Facultad de Ingenierfa Forestal, 
Universidad Austral de Chile, Valdivia, 1979. 

Smith, V.N. Selecci6n de irboles plus en Constituci6n, provincia de Maule. Actas V 
jornadas forestales. Asociaci6n Chilena de Ingenieros Forestales. Los Angeles, 
Chile, 1969. 

Aparico, J. and Siri, A. Estudio de variacion en plintulas de Pinus radiata. D. Don. 
en Chile. Tesis Facultad de Agrononrfa, Escuela de Ingenierfa Forestal, Universidad 
de Chile, Santiago, 1970. 

Garrido, F. Ibarra, M. Steinmetz, J. and Seron, J. Variacion de poblaciones naturales 
de RauH (Nothofagus alpina (Poepp. et Endl.) Oerst). Revisi6n bibliograf ica. 
CONAF-FAO. FO DP/CHI/ 76/003 . Documento de trabajo No. 28, Santiago de Chile, 1979. 

- 237 - 

Country statement: COLOMBIA 


1.1 Area of country: 1 138 914 km 2 

1.2 Location: longitude 7404'41'; latitude 435'56" 

1.3 Population: 28 000 000 

1.4 Main climatic and vegetative zones: 

There are various areas in the country; they include humid and very humid tropical 
forest on the Pacific coast and in Amazonia; dry tropical forest on the Atlantic 
coast and in the east of the country; and premontane and montane forests in the 
Andean area. 


2.1 Area of forests: 36 426 000 ha (1977) 

2.2 Proportion of land under forest: 38 percent 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

To rationally manage forest areas, which are classed as productive, productive- 
protective and protective. To encourage watershed management and conservation in 
the Andean region. To develop, through support services, such aspects as: 
forestation, protection, technical assistance, research and marketing of products. 

2.4 Legislation available to implement policy? YES 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Pulp, paper, cardboard - sawn wood - veneers, triplex, particle board, fibreboard - 
roundwood - posts, poles, fencing, cross-pieces, charcoal, rubber, balata, balsam of 
Tolu, essential oils, gum, etc. 

2.7 Staff State Others 

2.7.1 Professional: 270 650 

2.7.2 Subprofessional (with diploma 

or certificate of training): 30 90 


3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1978: 
80 000 ha 

3.1.2 Planned annual target/reforestation: 
40 000 ha /year 

3.2 Organization and administration of plantig schemes 

3.2.1 State Forest Services: 20 percent 

3.2.2 Others (specify): 

Forest corporations: 30 percent 
Private companies: 50 percent 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuelwood, protection, etc.) 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 238 - 

3.3.1 Indigenous species: 

Product/ Net Area 
Purpose Species (ha) I/ 



Cabinet- Tabebuia 
making rosea 250 

Cabinet* Alnus 
making jorullensis 500 

Cabinet- Cordia 
making alliodora 650 

Cabinet- Cariniana 
making pyriformis 700 

Introduced species: 



Mean annual increment (with- 

Product/ Net area 

Rotation out bark) at the end of the 

Purpose Species (ha) I/ 

years rotation (mVha/year) 

Pulp Pinus petula 25 000 

15-20 20m 3 /ha/year 

Roundwood Eucalyptus 


globuius 18 000 

8-15 22m J /ha/year 

Pulp Cupressus 


lusitanica 21 000 

15-25 18tn /ha/year 

Pulp Pinus 


radiata 3 000 

15 15m J /ha/year 

Construction Tectona 


grand is 1 000 

40 15m /ha/year 

Round wood Eucalyptus 


spp. 4 000 

12-15 ISin/ha/year 

Other species 6 000 





Service and control of seed supply 


Have seed zones been defined for indigenous species (i.e. is it common to restrict 

the use of seed for afforestation/reforestation to the particular zone in which it 

was collected)? YES 


Is there a national seed certification system? YES 


Are there facilities for storing seed at 

controlled temperatures? YES 


Does the supply of seed cover the demand 

for the species listed in 3.3.1? NO 

3.3.2? NO 




Does the country have an official tree improvement programme? YES 


If yes, list species involved: 

Pinus patula 
Cupressus lusitanica 
Pinus oocarpa 
Pinus caribaea 
Cordia alliodora 

Pinus kesiya 
Eucalyptus globuius 
Eucalyptus grandis 
Gmelina arborea 
Cariniana pyriformis 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
~ firebreaks, buildings and other areas not covered by trees. 

- 239 - 

5.1.2 Most important characters to be improved or bred for: 

Upward growth Form of the trunk 

Specific weight Angle and thickness of branches 

Fibre length Crown size 

Ease of pruning 

5.2 Brief outline of improvement methods already applied: 

5.2.1 Species/provenance trials (indicate in parentheses the number of provenances being 

Cupressus lusitanica (20) Eucalyptus globulus (32) 
Pinus caribaea (14) Gmelina arborea (11) 

Pinus oocarpa (19) Cariniana pyriformis (4) 

Pinus kesiya (25) Tectona grandis (5) 

Pinus patula (?) Cordia alliodora (13) 

Ochroma lagopus (5) 

5.2.2 Area of seed stands in each of the main species: 

Cordia alliodora 
Ochroma lagopus 
Jacaranda copaia 

5.2.3 Plus trees of main species (indicate in parentheses the number of trees): 

Cupressus lusitanica (43) pre-selected species 
Pinus patula (20) 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 
Cupressus lusitanica (5 ha, 43 clones) 

5.2.5 Progeny testing? (species and area): 

Cupressus lusitanica 
Pinus patula 
Cariniana pyriformis 

5.2.6 Other methods of improvement (specify): 

Species Method 

Bombacopsis quinata Large cutting 

Tabebuia rosea Large cutting 

Cupressus lusitanica Lateral grafting 

Pinus patula Grafting 


Logros geneticos con Cupressus lusitanica a t raves de 6 afios de mejoramiento geneticc 
en Colombia. Carton de Colombia: Investigacion Forestal. 

Documentos de Trabajo del Proyecto sobre Investigaciones y Desarrollo Industrial 
Forestaies, INDERENA/PNUD/FAO/CONIF COL/74/005: 

Konig, A, & Melchior, G.H. Propagacion vegetativa en arboles forestaies; 
Konig, A. & Melchior, G.H. & Venegas T.L. Ensayos de procedencias con Pinus 


Konig, A. Ensayos de procedencia con Pinus oocarpa y Pinus kesiya en Colombia; 
Melchior, G.H. Bases y posibilidades del mejoramiento de arboles forestaies en 

- 240 - 

Country statement: COSTA RICA 

1*1 Area of country: 50 851 km 2 

1.2 Location: Longitude: 8234 f - 8559'; latitude: 802 f - 

1.3 Population: 2 125 620 

1.4 Main climatic and vegetative zones: 

Climatic zones: Valle Intermontano (Inter-montane valley), Vertiente Pacifica 
(Pacific slope) , Vertiente Atlantica (Atlantic slope) and the 

Vegetative zones: Tropical forest (dry-humid), sub-tropical forest (humid-very 
humid) , low montane forest (humid-very humid- rainy) , montane 
forest, sub-alpine forest. 


2.1 Area of forests: 2 088 200 ha 

2.2 Proportion of land under forest: 40.9% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

A. To transform productive forests into an efficient system for the supply of 
raw material to industry. 

B. To conserve protective forests. 

C. To encourage large-scale reforestation as a better land use measure. To 
strengthen the State institutions for forest planning and control. 

2.4 Legislation available to Implement policy? YES 

2.5 Ownership of forests: 

Under State control: 11.5% 

Private ownership: 35.0% 

Community ownership: 0.3% 
No effective contro: 

2.6 Principal forest products (for exemple, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Sixty-five percent of the country's forest production is saw-wood to meet internal 
needs, and some for furniture for export. The remainder is processed as plywood 
and particle board for export. A low percentage is used for the preparation of 
sleepers and souvenirs. 

2.7 Staff State 

2.7.1 Professional: 20 

2.7.2 Subprofessional (with diploma or 

certificate of training): 127 29 

2.8 Gross annual budget for forestry: US$ 2 264 655.40 

3.1 Areas 

3.1.1 Net total area - of plantations at the end of 1978: 810 ha 

3.1.2 Planned annual target /reforestation: 6 200 ha/year 

II The net area is the gross area of plantation, less the area occupied by roads* paths, 
firebreaks, buildings and other areas not covered by trees. 

- 241 - 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: 82 

Others (Specify): Private ownership: 92% 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuelwood, protection, etc.) 

3.3.1 Indigenous species: 






Saw-wood Cordla 


Saw-wood Bombacopsis 

Saw-wood Alnus 


3.3.2 Introduced species: 

Net area 
(ha) I/. 










Saw-wood Pinus 


Posts Eucalyptus 

Saw-wood Cupressus 



Saw-wood Tectona 


Net area 
(ha) I/ 












Mean annual increment 
(without bark) at the 
end of the rotation 



Mean annual Increment 
(without bark) at the 
end of the rotation 



Service and control of seed supply 

Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 
it was collected)? YES 

Is there a national seed certification system? YES 

Are there facilities for storing seed at controlled temperature? YES 

Does the supply of seed cover the demand for the species listed 

in 3.3.1? NO 
in 3.3.2? NO 

? NO 


5.1 Does / the country have an official tree improvement progra 


Please list the reference to tree Improvement in the country, in publications, 
report, etc. 

There is no literature on this specific subject. The only publications are a few 
short information booklets on identification of seed trees, published by the 
Forest Service of Costa Rica. 

I/ The net area is the gross area of plantations, less the area occupied by road*, paths, 
firebreaks, buildings and other areas not overed by trees. 

- 242 - 

Country Statement: CUBA 


1.1 Area of country: 110 922 km 2 

1.2 Location: longitude: 74 and 85; latitude: 19 and 24 

1.3 Population: 9 405 000 


2.1 Area of forests: 1 691 131 ha 

2.2 Proportion of land under forest: 15t 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

Reforestation of all potential forest areas to cover an estimated 27 percent 
of the country. 

Well-managed forests for the purposes of production, protection of the environment 
and wildlife, and recreation. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: Under State control: 100% 


4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 

it was collected)? YES 

4.1.2 Is there a national seed certification system? YES 

4.1.3 Are there facilitaties for storing seed at controlled temperatures? YES 

4.1.4 Does the supply of seed cover the demand for the species listed 

in 3.3.1? YES 
in 3.3.2? YES 


5.1 Does the country have an official tree improvement programme? YES 

5.1.1 If yes, list species involved: 

Pinus caribaea var. caribaea Hibiscus elatus 

P. tropicalis Cedrela odor at a 
P. cubensis 
P. maestrensis 

5.1.2 Most important characters to be improved or bred for: 


Stem straightness 


5.2 Brief ouline of improvement methods already applied: 

5.2.1 Species /provenance trials (indicate in parantheses the number of provenances being 
tested) : 

Pinus caribea var. caribea (13) Hibiscus elatus (23) 
P. tropicalis (6) Cedrela odorata (13) 

P. cubensis (10) Tectona grandis (16) 

P. maestrensis (14) Eucalyptus saligna (16) 

Svietenia macrophylla (8) P. caribaea var. hondurensis (9) 

- 243 - 

5.2.2 Area of seed in each of the main species: 

Pinus caribaea var. caribaea (1 280 ha) P. maestrensis (147 ha) 
P. tropicalis (175 ha) Hibiscus elatus (130 ha) 

P. cubensis (572 ha) Cedrela odorata (19 ha) 

5.2.3 Plus trees of main species (indicate in parentheses the number of trees): 

P. caribaea var. caribaea (270) P. maestrensis (72) 

P. tropicalis (120) Hibiscus elatus (90) 

P. cubensis (150) Cedrela odorata (220) 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 

Pinus caribaea var. caribaea (200 ha) (108 cl.) 
P. cubensis (45 ha) (98 cl.) 
Hibiscus elatus (13 ha) ,(48 cl.) 

5.2.5 Progeny testing? (species and area): 

Pinus caribaea var. caribaea (27,5 ha) P. maestrensis (7 ha) 
P. tropicalis (21,5 ha) Hibiscus elatus (3 ha) 

P. cubensis (6 ha) Cedrela odorata (15 ha) 

Species Method % of success 

Pinus caribaea var. caribaea Grafting 80 

P. cubensis " 65 

P. maestrensis " 70 

Hibiscus elatus " 90 

Cedrela odorata " 85 


Please list the references to tree improvement in the country, in publications, 
reports, etc. 

7.1 Betancourt, A. and Gonzales A. (1972) "Trabajos realizados en Cuba sobre mejora- 
miento gentico de P. carbaea Mor. var. caribaea". In: Special Papers from Cuba 
at the Seventh World Forestry Congress, p. 7-26. 

7.2 Gonzales, A. (1974) "Algunos aspectos sobre mejoramiento selective de Srboles" 
(mimeographed) GIF, La Habana, 19 pp. 

7.3 Thomsen, J. (1974) "Gentica forestal y mane jo de semillas fores tales (ditto) 
GIF. La Habana, 246 pp. 

7.4 Zayas, A. and Bur ley, J. (1973) "Estudio de progenies de P. caribaea var. caribaea" 
Cajalbana, in Cuba, 7 pp. 

- 244 - 

Country Statement: DOMINICAN REPUBLIC 

1.1 Area of country: 48 442 km2 

1.3 Population: 5 200 000 

1.4 Main climatic and vegetative cones: 
Low montane humid forest 

Sub- tropical dry forest 


2.1 Area of forests: 1 100 000 ha 

2.2 Proportion of land under forest: 23% 

2.3 Does the country have a written statement of National Forest Policy? NO 

2.4 Legislation available to implement policy? NO 

2.5 Ownership of forests: 
Under State control: 701 
Private ownership: 20% 

Community ownership: - No effective control: 10% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Fuel wood 

Piles for fencing 

Poles for power lines 

2.7 Staff State Others 

2.7.1 Professional: 2 4 

2.7.2 Subprofessional (with diploma or 

certificate of training): 30 5 

2.8 Gross annual budget for forestry: US$ 3 000 000 


3.1 Areas 

3.1.1 Net total area If of plantations at the end of 1978: 
2 000 ha 

3.1.2 Planned annual target/reforestation: 
1 200 ha/year 

3.2 Organisation and administration of planting schemes 

3.2.1 State Forest Services: 95% Others (specify): 5 mining and private companies 

3.3 Principal productor purpose envisaged (for example, taw-wood, posts and stakes, 
pulwood, fuelwood, protection, etc.): 

3.3.1 Indigenous species: 

Mean annual increment (without bark) 

Product/ Net Area Rotation at the end of the rotation 

Purpose Species (ha) I/ (years) (m3/ha/year) 

Protection Pinus 

occidental is 1 000 30 

II The net area is the gross area of plantations, less the area occupied by roads, paths, 
*" firebreaks, buildings and other areas not covered by trees. 

- 245 - 

3.3.2 Introduced species: 


Protection and 




robust a 

Net area 
(ha) I/ 




Mean annual increment 
(without bark) at the 
end of the rotatiqn 


4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 

it was collected)? NO 

4.1.2 Is there a national seed certification system? NO 

4.1.3 Are there facilities for storing seed at controlled temperatures? YES 

4.1.4 Does the supply of seed cover the demand for the specied listed 

in 3.3.1? 
in 3.3.2? 



5.1 Does the country have an official tree improvement programme? NO 

- 246 - 

Country Statement: ECUADOR 


1.1 Area of country: 2 745.4 km 2 

1.2 Location: longitude 7830 f ; latitude: 00 f 

1.3 Population: 7 000 000 

1.4 Main climatic and vegetative zones: 

From tropical desert to perennial snowfields 


2.1 Area of forests: 17 734 000 ha 

2.2 Proportion of land under forest: 64.45% 

2.3 Does the country have a written statement on National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

Domestic utilisation of forest raw materials and conservation of renewable 
natural resources. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: 

Under State control: 12.41% 

Private ownership: 0.12% 

Community ownership: 0.02% 

No effective control: 87.45% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood) : 

Saw-wood; logs for triplex and plywood sleepers; posts; stakes; fuelwood and 
charcoal . 

2.7 Staff State Others 

2.7.1 Professional 53 20 

2.7.2 Subprofesslonal (with diploma 

or certificate of training) 114 33 

2.8 Gross annual budget for forestry: US$ 600 000 


3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1976: 25 425 ha 

3.1.2 Planned annual tar get reforest at ion: 5 000 ha/year 

3.2 Organization and administration of planting schemes 

3.2.1 State Forest Services: 70.0% 
Others (specify): 30.0% 

3.3 Principal product or purpose envisaged (for example, saw wood, posts and stakes, 
pulpwood, fuelwood, protection, etc.): 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 247 - 


Indigenous species: 


Species Net area 


Mean annual 


(ha I/) 



(without bark) 

at the end of 

the rotation 

(m^ /ha/year) 

Saw-wood and 

Eucalyptus glo- 17 405 






Pinus radiata 5 177 



Saw-wood and 

Tectona grandis 646 




Saw-wood and 

Eucalyptus 156 





Saw-wood and 

Eucalyptus 132 





Various uses 

Various species 129 




Introduced species: 


Species Net area 


Mean annual 

(ha) I/ 



(without bark) 

at the end of 


Saw-wood and 

Cordia alliodora 1 408 





Prosopis juliflora 173 



Saw-wood and 

Centrolobium 39 






Pseudosamanea 38 





Various species 122 






Service and control 

of seed supply 


Have seed zones been defined for indigenous species 

(i.e. is it common to restrict 

the use of seed for 

afforestation/reforestation to 

the particular 

zone in which it 

was collected)? NO 

- It is collected for local reforestation. 


Is there a national 

seed certification system? NO 


Are there facilities 

for storing seed at controlled 




Does the supply of seed cover the demand for the species listed in 3.3.1? YES 

in 3.3.2? For Eucalyptus globulus and Tectona fcrandis. 




Does the country have an official tree improvement 

programme? NO 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths 
firebreaks, buildings and other areas not covered by trees. 


Country Statement: GUATEMALA 

1.1 Area of country: 108 900 km 2 

1.3 Population: 7 045 800 

1.4 Main climatic and vegetative zones: 

a) South coast: Tropical and sub-tropical forests 

b) Western high plateau: Cold, humid and very himid, mixed and coniferous forests. 

c) Central region: Conifers in the western cold humid zone, tropical and dry 
sub-tropical forests in the east. 

d) Eastern high plateau: Cold, dry, conifers and mixed forests. 

e) Northern intermediate: Humid and very humid tropical, conifers and hard woods. 

f) North (PetSn): Tropical forests of valuable species, very humid. 


2.1 Area of forests: 3 610 000 ha 

2.2 Proportion of land under forest: 39.11 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

a) To conserve the existing forest area; 

b) To extend the forest area through afforestation and managed artificial 
reforestation; and 

c) To encourage the rise of forest industries. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: Under State control: 100% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Saw-wood (palo bianco, mahogany, cedar, pine); 
logs for veneering, for posts and for fuel wood. 

2.7 Gross annual budget for forestry: US$ 11 000 000 


3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1978: 3 610 000 ha 

3.1.2 Planned annual target reforestation: 12 000 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: 100Z 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuelwood, protection, etc.): 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 249 - 

3.3.1 Indigenous species: 

Product /Purpose 

3.3.2 Introduced species: 

Product /Purpose 




Net area 
(ha) I/ 


Mean annual increment 
(without bark) at the 
end of the rotation 
(m 3 /ha/year) 

3 000 
9 000 




Mean annual increment 
(without bark) at the 

Net area Rotation end of the rotation 
(ha) I/ (years) (m3/ha/year) 

Recently introduced 

Recently introduced 


4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 

it was collected? YES - Government Decree of 8 March 1979. 

4.1.2 Is there a national seed certification system? YES, BANSEFOR 

4.1.3 Are there facilities for storing seed at controlled temperatures? YES, BANSEFOR 

4.1.4 Does the supply of seed cover the demand for the species listed 

in 3.3.1? 
in 3.3.2? 






Does the country have an official tree improvement programme? NO 

Area of seed stands in each of the main species: 


1 295 ha 
200 ha 

5.2.3 Plus trees of main species (indicate in parentheses the number of trees): 

Conifers (155) 
Hardwoods - 

5.2.4 Seed orchard (species, area and number of clones or mother trees)* 

Leucaena leucocephala 
Sesbania aculeata 

3 ha (1 000) 
1 ha (100) 

1J The net area is the gross area of plantations, less the area occupied by roads, 
paths, firebreaks, buildings and other areas not covered by trees. 

- 250 - 

Country Statement: HONDURAS 


1.1 Area of country: 113 000 km 2 

1.2 Location: longitude: 85<>; latitude: 14 

1.3 Population: 4 million 

1.4 Main climatic and vegetative zones: Tropical 


2.1 Area of forests: 7.4 million ha 

2.2 Proportion of land under forest: 65% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

To preserve for the nation the inestimable flora, fauna and soil resources 

in forest areas; 

To ensure the protection and improvement of the above; 

To rationalize the utilization, industrialization and marketing of forest products. 

2.4 Legislation available to implement policy? YES 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Saw-wood, veneers and sleepers, fuelwood. 

2.7 Staff State Others 
Professional: 76 

Subprofessional (with diplc 

or certiflcat of training): 196 

2.8 Gross annual budget for forestry: US$ 80 million 


3.1 Areas 

3.1.2 Planned annual target area of afforestation/reforestation: 12 600 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: 100% 


4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 

it was collected)? NO 

4.1.2 Is there a national seed certification system? YES 

4.1.3 Are there facilities for storing seed at controlled temperatures? YES 

4.1.4 Does the supply of seed cover the demand for the species listed 
in 3.3.1? YES 

in 3.3.2? Very little work 


5.1 Does the country have an official tree improvement programme? YES 

- 251 - 

5.1.1 If yes, list species Involved: 

Pinus oocarpa Llquidambar styraciflua 

Pinus carlbaea Dldymopanax morototoni 

Pinus pseudostrobus Quercus sp. 

Cordla alliodora Leucaena leucocephala 

5.1.2 Most important characters to be Improved or bred for: 

Shape of stem Basic density 

Resin production 
Fibre length 

5.2 Brief outline of improvement methods already applied: 

5.2.1 Species /provenance trials (Indicate in parentheses the number of provenances 
being tested) : 

Pinus oocarpa (29) Pinus pseudostrobus (18) 

Pinus caribaea (26) Cordia alliodora (6) 

5.2.2 Area of seed stands in each of the main species: 
Not yet decided. 

5.2.3 Plus trees of main species (indicate in parentheses the number of trees): 
Selection has not yet been completed. 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 
The first one will be established this year; there are no data as yet. 

5.2.5 Progeny testing (species and area): 
To be started in 1980. 


Please list the references to tree improvement in the country, in publications, 
reports, etc. 

We are trying to publish some experiences. 

~ 252 - 

Country Statement: NICARAGUA 


1.1 Area of country: 130 000 km 2 

1.2 Location: longitude: 1045 f - 1505^ latitude: 8330 f - 8720 f 

1.3 Population: 2 500 000 

1.4 Main climatic and vegetative zones: 

Tropical dry forest. Tropical humid forest. Montane and pine wood savanna. 
Tropical rain forest, monsoon forest, tropical savanna and sub-tropical montane 


2.1 Area of forests: 4 550 000 ha 

2.2 Proportion of land under forest: 35Z 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

Nationalization and rational use of forests. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: 

Under state control: 20% 
No effective control: 80% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Saw-wood: 1 000 000 m3/year 

2.7 Staff State Others 
Professional : 7 3 

Subprofessional (with diploma 

or certif cate of training) : 12 3 

2.8 Gross annual budget for forestry: 
US$ 1.5 x 10 6 


3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1978: 
1500 ha 

3.1.2 Planned annual target /reforestation: 5 000 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: 100% 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuelwood, protection, etc.): 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 253 - 

3.3.1 Indigenous species: 




Timber for pulping Pinus 


Fuelwood Leucaena sp. 
Introduced species: 


Protection and 


Net Area 
(ha) I/ 

3 000 

Net Area 
(ha) I/ 

Eucalyptus 1 000 
Camaldulensis 500 




Mean annual increment 
(without bark) at the 
end of the rotation 
(m 3 /ha/year) 


Mean annual Increment 
(without bark) at the 
end of the rotation 
(m 3 /ha/year) 



4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is It common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which 

it was collected)? NO 

4.1.2 Is there a national seed certification system? NO 

4.1.3 Are there facilties for storing seed at controlled temperatures? NO 

4.1.4 Does the supply of seed cover the demand for the species listed 

in 3.3.1? 
in 3.3.2? 



5.1 Does the country have an official tree improvement programme? NO 

I/ The net area is the gross area of plantations, less the area occupied by roads, 
~ paths, firebreaks, buildings and other areas not covered by trees. 

- 254 - 

Country Statement: PANAMA 


1.1 Area of country: 77 082 km 2 

1.2 Location: Longitude: 779' - 833'; latitude: 712' - 938' N 

1.3 Population 1 881 400 

1.4 Main climatic and vegetative zones: 

Dry tropical forest, humid tropical forest, very humid tropical forest, very humid 
premontane forest, premontane rainforest, very humid low montane forest. 


2.1 Area of forests: 3 225 900 ha 

2.2 Proportion of land under forest: 50% 

2.3 Doest the country have a written statement of National Forest Policy? YES 

2.3.1 If a National Policy exists, what are the main objectives stated in it? 

To reduce importation by developing local industries, promoting local products, 
with the intention of remedying shortages in local supplies of the raw material 
necessary for production of wood and wood products. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: Under state control: 100% 

2.6 Principal forest products (for example, saw-wool, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Furniture, posts, beams, saw-wood, logs for sleepers, plywood, lumber. 


3.1 Areas 

3.1.2 Planned annual target area of afforestation/reforestation: 10 000 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: 100% 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuel wood, protection, etc.): 

3*3.1 Indigenous species: 

Mean annual increment (without bark) 


and beams 

Net Area Rotation 
Species (ha) I/ (years) 

at the end of the rotation 
(m 3 /ha/year) 










I/ The net area in the gross area of plantations, less the area occupied by roads, paths 
firebreaks, buildings and other areas not covered by trees* 

- 255 - 

3.3.2 Introduced species: 







Mean annual increment (without bark) 
Net Area Rotation at the end of the rotation 
(ha) I/ (years) (m3/ha/year) 

Posts, wood Tectona 



Reforestation Pinus 


2 400 


cadamba 1 . 5 




Service and control of seed supply 

Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which it 
was collected)? YES 

Is there a national seed certification system? YES 

Are there facilities for storing seed at controlled temperatures? YES 

Does the supply of seed cover the demand for the species listed in 3.3.1? 65% 
in 3.3.2? 50% 


Does the country have an official tree improvement programme? NO 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
"" firebreaks, buildings and other areas not covered by trees. 

- 256 - 

Country Statement: PARAGUAY 


1.1 Area of country: 406 752 km 2 

1.2 Location: Longitude: 54 - 63W; latitude: 19 - 28S 

1.3 Population: 3 500 000 

1.4 Main climatic and vegetative zones: 
Sub-tropical temperate clirnate 

Vegetation: thorny steppe and dry forest (western region) - HOLDRIDGE 
Humid forest (eastern region) - HOLDRIDGE 


2.1 Area of forests: 6 000 000 ha 

2.2 Proportion of land under forest: 45% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

a) Management plan; b) Reforestation plan; c) National parks plan; d) Five-year 
plan for industry. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: 
Under State control: 10% 
Private ownership: 90% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, 
veneer logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Sawn wood; veneer logs, logs for sleepers, posts, poles and stakes, fuelwood and 

2.7 Staff State Others 
Professional: 32 10 

Subprofessional (with diploma 

or certificate of training): 55 10 

2.8 Gross annual budget for Forestry: 
US$ 430 000 


3.1 Areas 

3.1.1 Net total area I/ of plantation at the end of 1978: 5 000 ha 

3.1.2 Planned annual target area of afforestation/reforestation: 1 500 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: - Others (specify): 100% (private) 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 257 - 

3*3.2 Introduced species: w . . 

Mean annual increment 

(without bark) at the 

Product/ Net Area Rotation end of the rotation 
Purpose Species (ha) J7 (years) (m3 /ha/year) 

Pulpwood Pinus elliottii 4 500 25 21.1 

Pulpvood Pinus taeda 30 25 21.3 

Saw- wood Araucaria angustifolia 3 30 15.8 

Pulpwood Cupressus japonica 2 25 9.2 

Note: Results in stands of 13 years, with 20 percent thinning at 9 years. 


4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it conmon to restrict 
the use of seed for afforestation/reforestation to the particular zone in which it 
was collected)? NO 

4.1.2 Is there a national seed certification system? NO 

4.1.3 Are there facilities for storing seed at controlled temperatures? YES 

4.1.4 Does the supply of seed cover the demand for the species listed in 3.3.1? NO 
in 3.3.2? NO 


5.1 Does the country have an official tree improvement programme? NO 

5.1.2 Most important characters to be improved or bred for: 

a) straight bole 

b) rapid growth 

c) physical-mechanical aptitudes 

d) resistance to pests, diseases and adverse climatic conditions 

5.2 Brief outline of improvement methods already applied: 

5.2.1 Species /provenance trials (indicate in parentheses the number of provenances 
being tested). Various exotic species of unknown specific origin have been 
introduced . 

Pinus elliottii (2) P. caribaea var. caribaea (2) 

P. taeda (2) P. caribaea var. hondurensis (1) 

Araucaria angustifolia (1) P. palustris (1) 
Cupressus japonica (1) Eucalyptus spp. (12 spp.) 

5.2.6 Other methods of improvement (specify): 

Note: In the second half of 1979, needs of various exotic species (conifers, and 
hardwoods) were introduced; collection of seeds of native species was also started, 
all of them tested but without final results. 


Note: We have no official reports on forest improvement, except for some publica- 
tions of partial results relating to the introduction of species, adaptability, etc. 

I/ H ntt area it the gross area of plantations, less the area occupied by roads, paths, 
~ firebreaks, buildings and other areas not covered by trees. 

- 258 - 

Country Statement: PERU 


1.1 Area of country: 1 285 215 km 2 

1.2 Location: Longitude: 70 - 815 latitude: - 18S 

1.3 Population: 20 000 000 

1*4 Main climatic and vegetative zones: 

Coastal region: dd-PT, ds-PT, dp-PT (ppx: 0-200 mm; T.x: 21 C) 
Mountain region: bb-MT, ee-MBT, bmh-MT (ppx: 250-1 400 mm; T.x: 14) 
Forest region: bh-T, bmh-T (ppx: 900-4 000 mm; T.x: 24 C) 


2.1 Area of forests: 74 120 000 ha 

2.2 Proportion of land under forest: 60% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

To meet the national demand for wood and forest products. 

To earmark land suitable for forests through reforestation to promote social and 
economic development. 

To lay the basis for optimum utilization of forests and wildlife resources. 

2.5 Ownership of forests: 
Under State control: 99.87. 
Private ownership: 0.047. 
Community ownership: 0.16% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer 
logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Saw-wood, veneer logs, plywood, parquet, sleepers, posts, timbers for mines, fuel- 
wood, charcoal. 

2.7 Staff State Others 
Professional: 240 

Subprofessional (with diploma or 

certificate of training): 130 

2.8 Gross annual budget for forestry: 
US$ 6 000 000 approximately 


3.1 Areas 

3.1.1 Net total area y of plantations at the end of 1978: 120 000 ha 

3.1.2 Planned annual target/reforestation: 15 000 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State forest Services: 98% 

Others (specify): Private ownership: 2% 

I/ The net area in the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 259 - 



product or purpose envisaged 'for example, saw-wood, posts and stakes, 


fuelwood, protection, etc.) 





Net Area 


Species (ha) I/ 


Cedrela sp. ) 


Swietenia ) 1 800 


Cedrelinga ) 


Prosopis sp. ) 


Tecoma sp. ) 4 000 


Lexopterygium ) 




Mean annual increment (without bark) 


Net Area Rotation at the end of the rotation 


Species (ha) I/ (ha) I/ (m3/ha/year) 






108 000 20 18-20 



Sawn wood 

Pinus radiata 3 000 20 14-18 









5.2 , 

Sawn wood 

Cupressus sp ) 
Pinus caribaea ) 
Eucalyptus ) 
grandis ) 
E. saligna ) 





Service and control of seed supply 

Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which it 
was collected)? YES 

Is there a national seed certification system? YES 

Are there facilities for storing seed at controlled temperatures? YES 

Does the supply of seed cover the demand for the species listed in 3.3.1? YES 
in 3.3.2? NO 


Does the country have an official tree improvement progr 

? YES 

If yes, list species involved: 
Pinus radiata 
? greggii 
P. patula 

P. elliottii 
Eucalyptus globulus 
E. grandis 

Most important characters to be improved or bred for: 
Rapid growth 
Straight trunk 

Brief outline of improvement methods already applied: 

Area of seed stand in each of the main species: Pinus radiata (20 ha) 

I/ The net area is the gross area of plantations, less the area occupied by roads, 
~ paths, firebreaks, buildings and other areas not covered by trees. 

- 260 - 

Country Statement: URUGUAY 


1.1 Area of country: 187 000 km* 

1.2 Location: longitude: 54-58W; latitude: 30-35S 

1.3 Population: 3 000 000 

1.4 Main climatic and vegative zones: 

Due to its small size, the country does not have marked climatic or vegetative 
variations; the climate is mild temperate and the vegetation mainly grasslands. 


2.1 Area of forests: 746 000 ha 

2.2 Proportion of land under forest: 3.97. 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

Meeting the internal demand for forest products (sawn wood, pulp and paper, 
fuel-wood, lumber, etc.) 

Export of production surpluses 
Concentration of forest stands 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: 
Under State control: 3% 
Private ownership: 97% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer 
logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Fuelwood : 1 200 000 n>3 

Saw and board wood : 150 000 m^ 

Pulpwood : 115 000 to? 

Veneer wood : 12 000 m 3 

2.7 Staff State Others 
Professional: 32 60 

Subprofessional (with diploma or 

certificate of training): 8 20 


3.1 Areas 

3.1.1 Net total area I/ of plantations at the end of 1978: 157 500 ha 

3.1.2 Planned annual target area of afforestion/reforestation: 5 000 ha/year 

3.2 Organization and administration of planting schemes 
3.2.1 State Forest Services: 25% 

Others (specif iy): Private ownership: 75% 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuel wood, protection, etc.): 

I/ The net area is the gross area of plantations, leas the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by tree\ 

- 261 - 

3.3.2 Introduced species: 









Mean annual increment (without bark) 


Net area 


at the end of the rotation , 


Species (ha) \J 




P. elliottii, 

pulp and 

P. taeda, P. 25 000 








umbel lata, E. 


camaldulensis, 110 000 




E. globulus, 

E. grandis 

Crates and 


boxes, saw- 

8 000 





Service and control of seed supply 

Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which it 
was collected)? NO 

Is there a national seed certification system? 

Being established. 

Are there facilities for storing seed at controlled temperatures? NO 

Does the supply of seed cover the demand for the species listed 

in 3.3.1? 

in 3.3.2? NO 


Does the country have an official tree improvement programme? YES 
If yes, list species involved: 

Pinus elliottee, P. taeda 
P. patula, P. pinaster, 
P. radiata 
Eucalyptus grandis 

E. globulus, E. smithii 
E. maidenii, E. resinifera, 
E. Bosistoana, E. umbel lata 
Salix sp., Populus sp. 
Platanus occidentalis 

Most important characters to be improved or bred for: 

Growth Resistance to climatic factors 

Conformation and disease and pests 

Timber quality 

Brief outline of improvement methods already applied: 

Species/provenance trials (indicate in parentheses the number of provenances 
being tested): 

Pinus taeda (11) Eucalyptus umbellata (15) 

Pinus elliottii (5) Eucalyptus maidenii (3) 

Pinus patula (3) 
Eucalyptus grandis (15) 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
~ firebreaks, buildings and other areas not covered by trees. 

- 262 - 

5.2.2 Area of seed stands in each of the main species: 

Pinus taeda (20 ha) Eucalyptus grandis (1,5 ha) 

Pinus elliottii (15 ha) 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 

Pinus taeda: 4 ha, 12 mother trees 
Pinus elliottii: 4 ha, 19 mother trees 

5.2.5 Progeny testing (species and area): 
Eucalyptus grandis: 0.2 ha 

5.2.6 Other methods of improvement (specify): 

Clonal selections are being made on Salix sp., Populus sp. and Platanus occidentalis 

Species Method % of success 

Salis sp., Populus sp. Untreated cutting 90-100% 

Platanus occidentalis Untreated cutting 85-95% 


Please list the references to tree improvement in the country, in publications 
reports, etc. 

Krall, J. adaptabilidad de confferas de Norteamerica plantadas en el Uruguay y su 
susceptibilidad a insectos y enfermedades. Bole tin del Departamento Forestal N 16. 

Krall, J. Populus: datos sobre fenologfa y crecimiento en el Uruguay. Boletfn del 
departament Forestal NO 17. 

Krall, J. Fundamentos para nuevas introduce iones de Eucalyptus en el Uruguay. 
Facultad de Agronomia, Boletfn N 17. 

- 263 - 

Country Statement: VENEZUELA 


1.1 Area of country: 913 050 km 2 

1.2 Location: Longitude: 5948 f - 73ll t 49"W; latitude: 043' - 1211'46 M N 

1.3 Population: 15 000 000 

1.4 Main climatic and vegetative zones: 

Dry tropical forest, humid tropical forest, very dry tropical forest, very humid 
premontane forest. 


2.1 Area of forests: 47 971 000 ha 

2.2 Proportion of land under forest: 53% 

2.3 Does the country have a written statement of National Forest Policy? YES 
2.3.1 If a National Policy exists, what are the main objectives stated in it? 

The objectives follow the major guidelines for the three classes of forest: 
protective, productive and the secondary functions such an recreation, research, 
education, etc. 

2.4 Legislation available to implement policy? YES 

2.5 Ownership of forests: 
Under State control: 82% 
Private ownership: 18% 

2.6 Principal forest products (for example, saw-wood, rubber, beeswax and honey, veneer 
logs, logs for sleepers, posts, piles and stakes, pulpwood): 

Saw-wood, veneer logs, particle board, sleepers, posts. 

2.7 Staff State Others 
Professional: 160 294 

Subprofessional (with diploma or 
certificate of training): 

2.8 Gross annual budget for forestry: 
US$ 20 000 000 


3.1 Areas 

3.1.1 Net total area \j of plantations at the end of 1978: 101 850 ha 

3.1.2 Planned annual target area of afforestation/ reforestation: 32 000 ha/year 

3.2 Organization and administration of planting schemes 

3.2.1 State Forest Services: 85% Others (specify): Combined State and Private 

ownership: 107. 
Private ownership: 5% 

3.3 Principal product or purpose envisaged (for example, saw-wood, posts and stakes, 
pulpwood, fuelwood, protection, etc.): 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 264 - 

3.3.1 Indigenous species: 



Mean annual increment (with- 

Net Area Rotation out bark) at the end of the 
(ha) I/ (years) rotation (m 3 /ha/year) 

Mean annual increment (with- 
out bark) at the end of the 
rotation (mVha/year) 


Saw-wood Anacardium excel sa,) 
Swietenia macro- ) 
phylla, Oedrela ) 
odorata, Cordia ) 2 800 
alliodora, Cordia ) 
apurensis, Tabebuia) 
rosea ) 

3.3.2 Introduced species: 

Product/ Net Area Rotation 

Purpose Species (ha) I/ (years) 

Pulp Eucalypts 3 000 6 

Pulp Caribbean Pine 101 850 12 
Saw-wood Teak 620 30 

Protection Pine, Eucalypts 
and others 


4.1 Service and control of seed supply 

4.1.1 Have seed zones been defined for indigenous species (i.e. is it common to restrict 
the use of seed for afforestation/reforestation to the particular zone in which it 
was collected)? YES 

4.1.2 Is there a national seed certification system? YES 

4.1.3 Are there facilities for storing seed at controlled temperatures? YES 

4.1.4 Does the supply of seed cover the demand for the species listed 
in 3.3.1? YES in 3.3.2? NO 


5.1 Does the country have an official tree improvement programme? 

In-depth tree breeding studies are done only in autonomous research organizations, 
or institutes: there is no State programme at national level. 

5.1.1 If yes, list species involved: 

Pinus caribaea var. hondurensis 
Bombacopsis quinata 

5.1.2 Most important characters to be improved or bred for: 

Straightness of stem, bifurcations, anomalies, height, wood density and volume 
(Caribbean pine). Buttresses, bifurcation in Bombacopsis quinata. 

5.2 Brief outline of improvement methods already applied: 

5.2.1 Species/provenance trials (indicate in parentheses the number of provenances being 

Genero Pinus (32 spp.) 
Indigenous hardwoods (10 spp.) 
Exotic hardwoods (25 spp.) 

I/ The net area is the gross area of plantations, less the area occupied by roads, paths, 
firebreaks, buildings and other areas not covered by trees. 

- 265 - 

5.2.2 Area of seed stands in each of the main species: 

Pinus caribaea var. hondurensis (520 ha) - in process of establishment (CVG-CONARE) 

5.2.3 Plus trees of main species (indicate in parentheses the number of trees): 
There are 40 Pinus caribaea trees selected as candidates 

5.2.4 Seed orchards (species, area and number of clones or mother trees): 

Bombacopsis quinata 
Tabebuia rosea 
Pinus caribaea 

5.2.5 Progeny testing (species and area): 

Bombacopsis quinata 
Tabebuia rosea 

5.2.6 Other methods of improvement (specify): 

Controlled and open pollination of Bombacopsis quinata 


Species Method % of success 

Bombacopsis quinata Grafting and cutting 90 

Cedrela odorata Grafting 

Tubebuia rosea Grafting 

Pinus caribaea Grafting, aerial shoots 30-50 


Please list the references to tree improvement in the country, in publications, 
reports, etc. 

Quijada and Gutierrez. Estudio sobre la propagacion vegetativa de especies 
forestales venezolanas. 

Mendez and Luis. Etapas preliminares para el establecimiento de un rodal semi Hero 
en Chaguaras Estado Monagas. 

Norman Smith. Seleccion de arboles en Cachipo para establecer un huerto semillero 
de Pinus caribaea. (Tree selection in Cachipo for the establishment of a Pinus 
caribaea seed orchard). 

Appendix VII 


1. Publications distributed to all participants 

a. Statistics, experimental design, selection of species/provenances 

Freese, F., 1978. Me*todos estadfsticos eleroentales para tecnicos forestales. 
Servicio Forestal, Departamento de Agricultura de los EE.UU. de America. 
Manual de Agricultura Num. 317. Centre Regional de Ayuda Tecnica, AID. 

Burley, J. & Wood, P.J., 1979. (Editors). Manual sobre investigaciones de especies 
y procedencias con referencia especial a los tr6picos. CFI Tropical Forestry 
Papers no. 10 & 10A (Edici6n Espafiola). Oxford, Reino Unido. 

b. Conservation of forest genetic resources 

Roche, L. , The met ho do logy of conservation of forest genetic resources* Report on 
a Pilot Study. FO:MISC/75/8. FAO/UNEP, Rome. 

FAO, 1977. Report of the Fourth Session of the FAO Panel of Experts on Forest 
Genetic Resources. FO:FGR/4/Rep. FAO, Rome. 

FAO, 1975. Forest genetic resources information No. 4 (Global programme for improved 
use of forest genetic resources). Forestry Occasional Paper 1975/1. FAO, Rome. 

FAO, 1978. Forest genetic resources information No. 8. Forestry Occasional Paper 
1978/2. FAO, Rome. 

c. Reports on conferences and technical consultations 

FAO, 1978. Genetics, Selected papers from the Third World Consultation on 
Improvement of Forest Trees. Unasylva 30 (119/120). FAO, Rome. 

d. Forest tree improvement 

Faulkner, R., 1975. Seed Orchards. Forestry Commission Bulletin No. 54. London, 
U.K. 1975. 

Koenig, A. & Melchior, G.H., 1978. Propagacion vegetativa en arboles forestales. 
FAO/PNUD/ COL/74/005, P.I.F. No. 9. INDERENA, Bogota. 

e. Collection of forest seeds for storage and treatment 

Greaves, A. y 1978. Descriptions of seed sources and collection for provenances of 
Pinus caribaea. Commonwealth Forestry Institute, Oxford. Tropical Forestry 
Papers No. 12. Oxford, U.K. 

Bonner, F.T., 1977. Equipment and Supplies for Collecting, Processing, Storing and 
Testing Forest Tree Seed. USDA Forest Service, General Technical Report SO-13. 
Southern Forest Expt. Station, New Orleans, Louisiana. 

Yeatroan, C.W. & Nieman, T.C., 1978. Safe tree climbing in forest management. 
Petawawa Forest Experiment Station, Chalk River, Ontario. Forestry Technical 
Report 24. 

FAO 1977. Manual I, sobre elcccion de rodales para la recolecci6n de seraillas 
forestales. INAFOR-BANSEFOR-FAO. FAO/TCP GUA-6/01-T, Guatemala. 

FAO, 1978. Manual II, para la recolecci6n de semi 11 as forestales. INAFOR-BANSEFOR- 
FAO/TCP GUA-6/01-T, Guatemala. 

- 267 - 

van Dijk, K. , Venegas Tovar, L. , & Melchior, G.H. , 1978. El suministro de serai 11 as 
como base de reforestaciones en Colombia. FAO/PNUD COL/74/005, P.l.F. No. 13. 
INDERENA, Bogota. 

Gordon, A.W. , 1979. Uso y abastecimiento de semillas forestales en Chile. FAO/PNUD, 
DP/CHI. 76/003, Documento de Trabajo No. 16. Santiago de Chile. 

f . Information on specific species 

g. Reforestation, forest plantations 

FAO, 1978. Establishment Techniques for Forest Plantations. FAO Forestry Paper No. 8. 
FAO, Rome. 

Navarro, G. & Molina, R.J., & Montero de B., J.L., 1975. Tecnicas de forestaci6n. 
ICONA, Monograffas No. 9. (2a edition). Madrid, Spain. 

Sommer, A. & Dow, T., 1978. Compilation of indicative growth and yield data on fast- 
growing exotic tree species planted in tropical and sub-tropical regions. FO: MISC/ 
78/11. FAO, Rome. 

Lanley, J.P. & Clement, J. Present and Future Forest and Plantation Areas in the 
Tropics. FO/MISC/79/1. FAO, Rome. 

h. General 

FAO, 1978. Publications on sale. FAO, Rome. 

2. Publications available for reference 

a. Statistics, experimental design, selection of species/provenances 

FAO, 1970. Majoramiento de arboles y biometra, Cuba. FAO: SF/CUB 3, Informe 
Teen i co No. 1. FAO, Rome. 

FAO, 1972. Genetica y biometra, Cuba. FAO:SF/CUB 3, Informe TScnico No. 3. FAO, 

b. Conservation of forest genetic resources 

FAO, 1973-78. Forest genetic resources information, Nos. 1-8. Forestry Occasional 
Paper. FAO, Rome. 

Lucas, G. & Synge, H. , 1978. (Editores). The IUCN Plant Red Data Book. IUCN/WWF, 
Morges, Switzerland. 

Frankel, O.K. & Bennett, E., 1970. (Editores). Genetic resources in plants: their 
exploration and conservation. IBP Handbook No. 11. Blackwell Scientific 
Publications, London, U.K. 

c. Reports on conferences and technical consultations 

Anon, 1978. Actas de la tercera consulta mundial FAO/IUFRO sobre la mejora de 
arboles forestales, I-III. CSIRO, Canberra, Australia. 

Burley, L. & Nikles, D.G., 1972-73. (Editores). Selection and breeding to improve 
some tropical conifers. Based on papers submitted to a Symposium of IUFRO Working 
Groups on Breeding of Tropical and sub-tropical Species and on Quantitative 
Genetics applied to forest trees, held in Gainesville, Florida, 1971. CFI, 
Oxford, U.K. 

Burley, J. & Nikles, D.G. Tropical provenance and progeny research and international 
cooperation. Proceedings of a joint Meeting of IUFRO Working Parties S.2.02.8 and 
82. 03.1, held in Nairobi, Kenya 1973. CFI. Oxford, U.K. 1973. 

- 268 - 

Nikles, D.G., Burley, J. & Barnes, R.D., 1978. Progress and problems of genetic 
improvement of tropical forest trees. Proceeding of a joint workshop of IUFRO 
Working Parties S2.02.8 and S2.03.1, held in Brisbane, Australia, 1977. CFI, 
Oxford, U.K. 

FAO, 1963. Proceedings of the World Consultation on Forest Genetics and Tree 
Improvement, Stockholm. FAO, Rome. 

FAO, 1969. Proceedings of the Second World Consultation on Forest Tree Improvement, 
Washington, FAO, Rome. 

FAO, 1964. FAO/IUFRO meeting on forest genetics. Unasylva 18 (2-3), Nos. 73-74. 
FAO, Rome. 

FAO, 1970. FAO/IUFRO Consultation on frest tree breeding. Unasylva 24 (2-3) 
Nos. 97-98. FAO, Rome. 

d. Forest tree improvement 

Wright, J.W. Genetics of Forest Tree Improvement. FAO Forestry and Forest Products 
Studies No. 16. FAO, Rome. 

Allard, L.W. (trad. J.L. Montoya). Principios de la mejora genetica de las plantas. 
Editorial Omega S.A., Barcelona. 

Baldwin, R.E. Genetica elemental. Editorial Limusa, Mexico D.F., 1976. 

Hartmann, H.T. & Kester, D.E., 197- . Propagacion de plantas: teorfa y practice. 
Editorial Continental, Mexico. 

Stern, K. & Roche, L., 1974. Genetics of Forest Ecosystems. Ecological Studies 
No. 6. Chapman & Hall Ltd., London, U.K. 

Dorman, K.W. , 1976. The Genetics and breeding of southern pines. USDA Agric. 
Handbook No. 471 USDA, Forest Service. Washington D.C. 

FAO, 1974. Report on the FAO /DAN I DA Training Course on Forest Tree Improvement, 
Kenya 1973. FAO, Rome. 

ADAA, 1978. Selected reference papers. International Training Course in Forest Tree 
Breeding, Canberra 1977. Australian Development Assistance Agency, Australia. 

Burdon, R., 1978. Mejoramiento genetico forestal en Chile. FAO/PNUD DP/CHI/76/003, 
Documento de Trabajo No. 11. Santiago de Chile. 

Melchior, G.H., 1977. Bases y posibilidades del mejoramiento genetico de arboles 
forestales en Colombia. FAO/PNUD COL/74/005, P.I.F. No. 4. INDERENA, Bogota. 

Melchior, G.H., 1977. Programa preliminar de un ensayo de procedencia de Cordia 
alliodora, Cupressus lusitanica y otras especies nativas y ex6ticas. FAO/PNUD 
COL/74/005, P.I.F. No. 7. INDERENA, Bogota". 

Shelbourne, C.J.A., 1969. Tree Breeding Methods. New Zealand Forest Service. 
Technical Development Paper No. 55. Wellington, N.Z. 

e. Collection of forest seeds for storage and treatment 

FAO, 1975* Report on the FAO /DAN I A Training Course on Forest Seed Collection and 
Handling, I-II, Thailand 1975. FAO, Rome. 

1STA, 197- . Reglas Internacionales para ensayos de semi lias. Librerfa Agrfcola, 
Madrid, Espafia. 

Catalan Bachiller, G. t 1977. Semi 11 as de Arboles y arbustos forestales. Min. Agric. 
(1CONA) Madrid. 

- 269 - 

Anon, 1974* Seeds of Woody Plants in the United States. USDA Agric. Handbook No. 450 
USDA, Forest Service. Washington D.C. 

FAO, 1975. Forest tree seed directory. FAO, Rome. 

1BPGR, 1976. Report of IBPGR Working Group on engineering, design and cost aspects 
of long-term seed storage facilities. International Board for Plant Genetic 
Resources. FAO, Rome. 

Lowman, B.J., 1975. Equipment for processing small seedlots. Catalog No. 7524-2505. 
USDA Forest Service, Missoula, Montana. 

FAO, 1968. Forest tree seed notes. I. Arid areas. II. Humid tropics. FAO Forestry 
Development Paper No. 5. FAO, Rome. 

Purtich, G.S., 1972.* Cone production in conifers. Canadian Forestry Service, 
BC - X - 65. Department of the Environment. Victoria, British Columbia. 

Sarvas, R. , 1962. Investigations on the flowering and seed crop of Pinus 
silvestris L. Com. Inst. For. Fenniae 53.4. Helsinki. 

Melchior, G.H. & Venegas T.L., 1978. Propuesta para asegurar el suministro de 
semillas de Eucalyptus globulus en calidad comercial y geneticamente mejoradas. 
FAO/PNUD COL/74/005, P.I.F. No. 14. INDERENA, Bogota. 

FAO, 1978. Centro de Semillas forestales, Paraguay. Estudio de Factabilidad 
Tecnica para su establecimiento, organizacion y funcionamiento. FAO/SFN/TCP 
6/PAR/02-T. Asuncion, Paraguay. 

Bramlett, D.L., 1978. e_t aj^. Cone analysis of Southern Pines: a Guidebook. USDA, 
Forest Service General Technical Report No. SE-13. Washington D.C. 

f . Information on specific species 

FAO, 1973/74/75. Annotated bibliographies on Pinus patula, Cupressus lusitanica, 
and Pinus elliottii. FAO, Rome. 

Carruyo, L.J., 1976. Bibliograf fas sobre Pinus caribaea (1938-72; 1973-75), I - II. 
Laboratorio Nacional de Productos Forestales, Publicaciones Tecnicas 76-01-02, 
01-03 76. Merida, Venezuela. 

CFI, Oxford, "Fast Growing Timber Trees of the Lowland Tropics 11 , numbers: 

1. Gmelina arborea (1968) 

2. Cedrela odorata (1968) 
3.' The araucarias (1968) 

4. Pinus merkusii (1968) 

5. Terminalia ivorensis (1971) 

6. Pinus caribaea (1973) 

CFI, Oxford, "Tropical Forestry Papers", numbers: 

7. Pinus patula (1975) 

8. Eucalyptus camaldulensis (bibliography) (1975) 

9. Pinus kesiya (1979) 
10. Agathis (1977) 

US Department of Agriculture, Forest Service; Research Papers: 

WO - 19 (1973): Genetics of loblolly pine 
WO - 20 (1974): Genetics of slash pine 
WO - 35 (1978): Genetics of douglas fir 

Gibson, I.A.S., 1975. Diseases of forest trees widely planted as exotics in the 
tropics and Southern hemisphere. I. Commonwealth Mycological Institute/ 
Commonwealth Forestry Institute, Oxford, U.K. 

- 270 - 

Gibson, I.A.S., 1979. Diseases of forest trees widely planted in the tropics and 
Southern hemisphere. II. The genus Pinus. CM1/CFI, Oxford, U.K. 

Greaves, A., 1979. Descriptions of seed sources and collections for provenances of 
Pinus oocarpa. CF1 Tropical Forestry Papers No. 13. Oxford, U.K. 

Rowland, P., 1979. Pericopsis elata (Af romorsia) . A summary of silvicultural 
knowledge. CFI, Oxford, U.K. 

Anon, 1976. Leucaena - promising forage and tree crop for the tropics. National 
Academy of Sciences, Washington D.C. 

Anon, 1976. Under-exploited tropical plants with promising economic value. National 
Academy of Sciences, Washington D.C. 

Venegas T.L., 1978. Distribucion de once especies forestales en Colombia. FAO/PNUD 
COL/ 74/005, P.I.F. No. 11. INDERENA, Bogota. 

Melchior, G.H., 1977. Propuesta para mejorar la produccion del Inchi por metodos 
silviculturales y geneticos. FAO/PNUD COL/74/005. P.I.F. No. 8. INDERENA, Bogota. 

FAO, 1971. Insectos que atacan a Pinus caribaea en el noreste de Nicaragua. 
FAO/SF/NIC 9, Informe Tecnico No.l. FAO, Rome. 

Ntima, O.O., 1971. Los Araucarias. Boll, 36-37, IFLAIC, Merida, Venezuela, 

FAO, 1980. Populus spp. Handbook on Plantation and Utilization. FAO, Rome (in press). 

FAO, 1980. Eucalyptus spp. Handbook on Plantation and Utilization. FAO, Rome 
(in press). 

g. Reforestation, forest plantations 

FAO, 1978. Choice of tree species. FAO Forestry Development Paper No. 13 (2nd Ed.). 
FAO, Rome. 

FAO, 1974. A manual on the planning of man-made forests. Fo:MISC/73/22. FAO, Rome. 

Paul, D.K., 1972. A handbook of nursery practice for Pinus caribaea var. hondurensis 
and other conifers in West Malaysia. FO:SF/MAL 12, Working Paper No. 19, 
Kuala Lumpur. 

FAO, 1960. Forest Plantation Practices in Latin America. Department of Forestry, FAO. 
FAO. Dev. Paper No. 15. FAO, Rome. 

Cozzo, D., 1976. Tecnologfa de la forestaci6n en Argentina y America Latina. Ed. 
Hemisferio Sur, Buenos Aires. 

FAO, 1975. Tree planting practices in African savannas. FAO Forestry Development 
Paper No. 19* FAO, Rome. 

FAO, 1977. Savanna Afforestation in Africa. FAO Forestry Paper No. 11. FAO, Rome 

FAO, 1978. Forestry for local community development. FAO Forestry Paper No. 7. 
FAO, Rome. 

FAO, 1979. Forestry for rural communities. FAO, Rome. 

FAO. 1967. FAO World Symposium on Manroade Forests and their Industrial Importance. 
Unasylva 21 (3-4), Nos. 86-87. FAO, Rome. 

Anon, 1978. Tropical Forest ecosystems: a state-of-knowledge report prepared by 
Unesco/UNEP/FAO. Natural Resources Research XIV, Unesco, Paris. 

- 271 - 





Con la presente se certified que 

participo en el Curso arriba mencionado 

celebrado en 

MMda, Venezuela 

del 14 de enero al 2 defebrero de 1980 

MMda, Venezuela 
2 de febrero de 1980